4730:
4075:
4725:{\displaystyle {\begin{aligned}&\mathbf {r} _{0}:=\mathbf {b} -\mathbf {Ax} _{0}\\&{\hbox{if }}\mathbf {r} _{0}{\text{ is sufficiently small, then return }}\mathbf {x} _{0}{\text{ as the result}}\\&\mathbf {p} _{0}:=\mathbf {r} _{0}\\&k:=0\\&{\text{repeat}}\\&\qquad \alpha _{k}:={\frac {\mathbf {r} _{k}^{\mathsf {T}}\mathbf {r} _{k}}{\mathbf {p} _{k}^{\mathsf {T}}\mathbf {Ap} _{k}}}\\&\qquad \mathbf {x} _{k+1}:=\mathbf {x} _{k}+\alpha _{k}\mathbf {p} _{k}\\&\qquad \mathbf {r} _{k+1}:=\mathbf {r} _{k}-\alpha _{k}\mathbf {Ap} _{k}\\&\qquad {\hbox{if }}\mathbf {r} _{k+1}{\text{ is sufficiently small, then exit loop}}\\&\qquad \beta _{k}:={\frac {\mathbf {r} _{k+1}^{\mathsf {T}}\mathbf {r} _{k+1}}{\mathbf {r} _{k}^{\mathsf {T}}\mathbf {r} _{k}}}\\&\qquad \mathbf {p} _{k+1}:=\mathbf {r} _{k+1}+\beta _{k}\mathbf {p} _{k}\\&\qquad k:=k+1\\&{\text{end repeat}}\\&{\text{return }}\mathbf {x} _{k+1}{\text{ as the result}}\end{aligned}}}
10627:
10168:
10622:{\displaystyle {\begin{aligned}\left\|\mathbf {e} _{k}\right\|_{\mathbf {A} }&=\min _{p\in \Pi _{k}^{*}}\left\|p(\mathbf {A} )\mathbf {e} _{0}\right\|_{\mathbf {A} }\\&\leq \min _{p\in \Pi _{k}^{*}}\,\max _{\lambda \in \sigma (\mathbf {A} )}|p(\lambda )|\ \left\|\mathbf {e} _{0}\right\|_{\mathbf {A} }\\&\leq 2\left({\frac {{\sqrt {\kappa (\mathbf {A} )}}-1}{{\sqrt {\kappa (\mathbf {A} )}}+1}}\right)^{k}\ \left\|\mathbf {e} _{0}\right\|_{\mathbf {A} }\\&\leq 2\exp \left({\frac {-2k}{\sqrt {\kappa (\mathbf {A} )}}}\right)\ \left\|\mathbf {e} _{0}\right\|_{\mathbf {A} }\,,\end{aligned}}}
14537:
9302:
20:
1565:
8117:
14291:
3271:
6803:
8807:
9039:
1281:
7829:
648:
8550:
2883:
2990:
14532:{\displaystyle W={\begin{bmatrix}t&{\sqrt {t}}&&&&\\{\sqrt {t}}&1+t&{\sqrt {t}}&&&\\&{\sqrt {t}}&1+t&{\sqrt {t}}&&\\&&{\sqrt {t}}&\ddots &\ddots &\\&&&\ddots &&\\&&&&&{\sqrt {t}}\\&&&&{\sqrt {t}}&1+t\end{bmatrix}},\quad b={\begin{bmatrix}1\\0\\\vdots \\0\end{bmatrix}}}
6542:
8577:
9297:{\displaystyle \alpha _{1}={\frac {\mathbf {r} _{1}^{\mathsf {T}}\mathbf {r} _{1}}{\mathbf {p} _{1}^{\mathsf {T}}\mathbf {Ap} _{1}}}\approx {\frac {{\begin{bmatrix}-0.2810&0.7492\end{bmatrix}}{\begin{bmatrix}-0.2810\\0.7492\end{bmatrix}}}{{\begin{bmatrix}-0.3511&0.7229\end{bmatrix}}{\begin{bmatrix}4&1\\1&3\end{bmatrix}}{\begin{bmatrix}-0.3511\\0.7229\end{bmatrix}}}}=0.4122.}
5773:
1560:{\displaystyle \mathbf {p} _{k}^{\mathsf {T}}\mathbf {b} =\mathbf {p} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {x} _{*}=\sum _{i=1}^{n}\alpha _{i}\mathbf {p} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{i}=\sum _{i=1}^{n}\alpha _{i}\left\langle \mathbf {p} _{k},\mathbf {p} _{i}\right\rangle _{\mathbf {A} }=\alpha _{k}\left\langle \mathbf {p} _{k},\mathbf {p} _{k}\right\rangle _{\mathbf {A} }}
6467:
8112:{\displaystyle \alpha _{0}={\frac {\mathbf {r} _{0}^{\mathsf {T}}\mathbf {r} _{0}}{\mathbf {p} _{0}^{\mathsf {T}}\mathbf {Ap} _{0}}}={\frac {{\begin{bmatrix}-8&-3\end{bmatrix}}{\begin{bmatrix}-8\\-3\end{bmatrix}}}{{\begin{bmatrix}-8&-3\end{bmatrix}}{\begin{bmatrix}4&1\\1&3\end{bmatrix}}{\begin{bmatrix}-8\\-3\end{bmatrix}}}}={\frac {73}{331}}\approx 0.2205}
488:
9487:
9003:
8305:
2593:
8330:
2685:
3266:{\displaystyle {\begin{aligned}f(\mathbf {x} _{k+1})&=f(\mathbf {x} _{k}+\alpha _{k}\mathbf {p} _{k})=:g(\alpha _{k})\\g'(\alpha _{k})&{\overset {!}{=}}0\quad \Rightarrow \quad \alpha _{k}={\frac {\mathbf {p} _{k}^{\mathsf {T}}(\mathbf {b} -\mathbf {Ax} _{k})}{\mathbf {p} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}}}\,.\end{aligned}}}
7783:
11055:
6798:{\displaystyle \mathbf {p} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}=(\mathbf {r} _{k}+\beta _{k-1}\mathbf {p} _{k-1})^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}={\frac {1}{\alpha _{k}}}\mathbf {r} _{k}^{\mathsf {T}}(\mathbf {r} _{k}-\mathbf {r} _{k+1})={\frac {1}{\alpha _{k}}}\mathbf {r} _{k}^{\mathsf {T}}\mathbf {r} _{k}}
8802:{\displaystyle \beta _{0}={\frac {\mathbf {r} _{1}^{\mathsf {T}}\mathbf {r} _{1}}{\mathbf {r} _{0}^{\mathsf {T}}\mathbf {r} _{0}}}\approx {\frac {{\begin{bmatrix}-0.2810&0.7492\end{bmatrix}}{\begin{bmatrix}-0.2810\\0.7492\end{bmatrix}}}{{\begin{bmatrix}-8&-3\end{bmatrix}}{\begin{bmatrix}-8\\-3\end{bmatrix}}}}=0.0088.}
2010:
5597:
6263:
1201:
13525:
7391:
3977:
13382:
In numerically challenging applications, sophisticated preconditioners are used, which may lead to variable preconditioning, changing between iterations. Even if the preconditioner is symmetric positive-definite on every iteration, the fact that it may change makes the arguments above invalid, and in
6059:
2432:
must be conjugate to each other. A practical way to enforce this is by requiring that the next search direction be built out of the current residual and all previous search directions. The conjugation constraint is an orthonormal-type constraint and hence the algorithm can be viewed as an example of
643:{\displaystyle \mathbf {u} ^{\mathsf {T}}\mathbf {A} \mathbf {v} =\langle \mathbf {u} ,\mathbf {v} \rangle _{\mathbf {A} }:=\langle \mathbf {A} \mathbf {u} ,\mathbf {v} \rangle =\langle \mathbf {u} ,\mathbf {A} ^{\mathsf {T}}\mathbf {v} \rangle =\langle \mathbf {u} ,\mathbf {A} \mathbf {v} \rangle .}
9533:
it produces the exact solution after a finite number of iterations, which is not larger than the size of the matrix. In practice, the exact solution is never obtained since the conjugate gradient method is unstable with respect to even small perturbations, e.g., most directions are not in practice
3281:
The above algorithm gives the most straightforward explanation of the conjugate gradient method. Seemingly, the algorithm as stated requires storage of all previous searching directions and residue vectors, as well as many matrix–vector multiplications, and thus can be computationally expensive.
9327:
11136:
If initialized randomly, the first stage of iterations is often the fastest, as the error is eliminated within the Krylov subspace that initially reflects a smaller effective condition number. The second stage of convergence is typically well defined by the theoretical convergence bound with
1680:
7569:
8834:
8135:
10814:
8545:{\displaystyle \mathbf {r} _{1}=\mathbf {r} _{0}-\alpha _{0}\mathbf {A} \mathbf {p} _{0}={\begin{bmatrix}-8\\-3\end{bmatrix}}-{\frac {73}{331}}{\begin{bmatrix}4&1\\1&3\end{bmatrix}}{\begin{bmatrix}-8\\-3\end{bmatrix}}\approx {\begin{bmatrix}-0.2810\\0.7492\end{bmatrix}}.}
13647:
11973:
2878:{\displaystyle \alpha _{k}={\frac {\mathbf {p} _{k}^{\mathsf {T}}(\mathbf {b} -\mathbf {Ax} _{k})}{\mathbf {p} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}}}={\frac {\mathbf {p} _{k}^{\mathsf {T}}\mathbf {r} _{k}}{\mathbf {p} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}}},}
2443:
11601:
7612:
370:
problems. Despite differences in their approaches, these derivations share a common topic—proving the orthogonality of the residuals and conjugacy of the search directions. These two properties are crucial to developing the well-known succinct formulation of the method.
12216:
10924:
12305:
12550:
would take more time/computational resources than solving the conjugate gradient algorithm itself. As an example, let's say that we are using a preconditioner coming from incomplete
Cholesky factorization. The resulting matrix is the lower triangular matrix
6246:
5768:{\displaystyle \alpha _{k}={\frac {\mathbf {r} _{k}^{\mathsf {T}}\mathbf {r} _{k}}{\mathbf {r} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}}}={\frac {\mathbf {r} _{k}^{\mathsf {T}}\mathbf {r} _{k}}{\mathbf {p} _{k}^{\mathsf {T}}\mathbf {Ap} _{k}}}}
12454:
has to be symmetric positive-definite and fixed, i.e., cannot change from iteration to iteration. If any of these assumptions on the preconditioner is violated, the behavior of the preconditioned conjugate gradient method may become unpredictable.
1895:
3793:
6462:{\displaystyle \mathbf {r} _{k+1}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}={\frac {1}{\alpha _{k}}}\mathbf {r} _{k+1}^{\mathsf {T}}(\mathbf {r} _{k}-\mathbf {r} _{k+1})=-{\frac {1}{\alpha _{k}}}\mathbf {r} _{k+1}^{\mathsf {T}}\mathbf {r} _{k+1}}
1057:
13393:
7258:
12861:
6145:
5859:
9971:
5945:
12782:
11860:
9482:{\displaystyle \mathbf {x} _{2}=\mathbf {x} _{1}+\alpha _{1}\mathbf {p} _{1}\approx {\begin{bmatrix}0.2356\\0.3384\end{bmatrix}}+0.4122{\begin{bmatrix}-0.3511\\0.7229\end{bmatrix}}={\begin{bmatrix}0.0909\\0.6364\end{bmatrix}}.}
12055:
11756:
5506:
5135:
11677:
11419:
5002:
11192:
and the spectral distribution of the error. In the last stage, the smallest attainable accuracy is reached and the convergence stalls or the method may even start diverging. In typical scientific computing applications in
2674:
2147:
12408:
4917:, i.e., can be used as a simple implementation of a restart of the conjugate gradient iterations. Restarts could slow down convergence, but may improve stability if the conjugate gradient method misbehaves, e.g., due to
1576:
8998:{\displaystyle \mathbf {p} _{1}=\mathbf {r} _{1}+\beta _{0}\mathbf {p} _{0}\approx {\begin{bmatrix}-0.2810\\0.7492\end{bmatrix}}+0.0088{\begin{bmatrix}-8\\-3\end{bmatrix}}={\begin{bmatrix}-0.3511\\0.7229\end{bmatrix}}.}
8300:{\displaystyle \mathbf {x} _{1}=\mathbf {x} _{0}+\alpha _{0}\mathbf {p} _{0}={\begin{bmatrix}2\\1\end{bmatrix}}+{\frac {73}{331}}{\begin{bmatrix}-8\\-3\end{bmatrix}}\approx {\begin{bmatrix}0.2356\\0.3384\end{bmatrix}}.}
3531:
908:
13146:
7474:
10708:
12310:
The
Cholesky decomposition of the preconditioner must be used to keep the symmetry (and positive definiteness) of the system. However, this decomposition does not need to be computed, and it is sufficient to know
452:
5424:
5350:
5202:
14704:
13720:
13191:
10160:
7455:
9570:
to the exact solution and may reach the required tolerance after a relatively small (compared to the problem size) number of iterations. The improvement is typically linear and its speed is determined by the
2588:{\displaystyle \mathbf {p} _{k}=\mathbf {r} _{k}-\sum _{i<k}{\frac {\mathbf {p} _{i}^{\mathsf {T}}\mathbf {A} \mathbf {r} _{k}}{\mathbf {p} _{i}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{i}}}\mathbf {p} _{i}}
11362:
5057:
3392:
2307:
27:
with optimal step size (in green) and conjugate vector (in red) for minimizing a quadratic function associated with a given linear system. Conjugate gradient, assuming exact arithmetic, converges in at most
13540:
11866:
9785:
13921:
804:
11503:
2074:
11115:
7778:{\displaystyle \mathbf {r} _{0}={\begin{bmatrix}1\\2\end{bmatrix}}-{\begin{bmatrix}4&1\\1&3\end{bmatrix}}{\begin{bmatrix}2\\1\end{bmatrix}}={\begin{bmatrix}-8\\-3\end{bmatrix}}=\mathbf {p} _{0}.}
12932:
12698:
12615:
11050:{\displaystyle {\frac {{\sqrt {\kappa (\mathbf {A} )}}-1}{{\sqrt {\kappa (\mathbf {A} )}}+1}}\approx 1-{\frac {2}{\sqrt {\kappa (\mathbf {A} )}}}\quad {\text{for}}\quad \kappa (\mathbf {A} )\gg 1\,.}
12119:
9831:
14212:
12968:
4012:
165:
10173:
4080:
2995:
12227:
11464:
10072:
13982:
1273:
12445:
11282:
11170:
9725:
14850:(2004). "Konrad Zuse und die ERMETH: Ein weltweiter Architektur-Vergleich" [Konrad Zuse and the ERMETH: A worldwide comparison of architectures]. In Hellige, Hans Dieter (ed.).
13277:
13246:
12652:
1237:
1049:
10897:
10696:
10661:
9862:
358:
The conjugate gradient method can be derived from several different perspectives, including specialization of the conjugate direction method for optimization, and variation of the
14824:
On the
Extension of the Davidon–Broyden Class of Rank One, Quasi-Newton Minimization Methods to an Infinite Dimensional Hilbert Space with Applications to Optimal Control Problems
2005:{\displaystyle f(\mathbf {x} )={\tfrac {1}{2}}\mathbf {x} ^{\mathsf {T}}\mathbf {A} \mathbf {x} -\mathbf {x} ^{\mathsf {T}}\mathbf {b} ,\qquad \mathbf {x} \in \mathbf {R} ^{n}\,.}
6505:
6155:
5901:
5560:
4882:
3788:
12526:
12341:
11245:
10863:
5256:
10101:
9568:
6534:
5930:
5589:
4915:
4814:
4781:
4063:
3728:
3677:
3648:
3615:
3586:
3472:
3421:
3338:
3309:
2915:
2430:
2388:
2339:
1806:
1777:
1016:
987:
342:
13783:
10837:
13372:
13350:
13328:
13299:
13215:
13084:
13058:
13036:
13014:
12992:
12885:
12571:
12548:
12494:
11304:
5224:
4847:
4034:
3750:
3699:
3443:
1196:{\displaystyle \mathbf {x} _{*}=\sum _{i=1}^{n}\alpha _{i}\mathbf {p} _{i}\Rightarrow \mathbf {A} \mathbf {x} _{*}=\sum _{i=1}^{n}\alpha _{i}\mathbf {A} \mathbf {p} _{i}.}
849:
739:
717:
695:
673:
477:
402:
313:
283:
238:
190:
14036:
14009:
5283:
2983:
2942:
1740:
9682:
9650:
9601:
216:
12089:
11197:
for matrices of large sizes, the conjugate gradient method uses a stopping criterion with a tolerance that terminates the iterations during the first or second stage.
10001:
2598:(see the picture at the top of the article for the effect of the conjugacy constraint on convergence). Following this direction, the next optimal location is given by
14286:
13520:{\displaystyle \beta _{k}:={\frac {\mathbf {r} _{k+1}^{\mathsf {T}}\left(\mathbf {z} _{k+1}-\mathbf {z} _{k}\right)}{\mathbf {r} _{k}^{\mathsf {T}}\mathbf {z} _{k}}}}
10917:
3557:
934:
11492:
7386:{\displaystyle \mathbf {A} \mathbf {x} ={\begin{bmatrix}4&1\\1&3\end{bmatrix}}{\begin{bmatrix}x_{1}\\x_{2}\end{bmatrix}}={\begin{bmatrix}1\\2\end{bmatrix}},}
15076:
Barrett, Richard; Berry, Michael; Chan, Tony F.; Demmel, James; Donato, June; Dongarra, Jack; Eijkhout, Victor; Pozo, Roldan; Romine, Charles; van der Vorst, Henk.
13656:, as it allows for variable preconditioning. The flexible version is also shown to be robust even if the preconditioner is not symmetric positive definite (SPD).
3972:{\displaystyle x_{k}=\mathrm {argmin} _{y\in \mathbb {R} ^{n}}{\left\{(x-y)^{\top }A(x-y):y\in \operatorname {span} \left\{b,Ab,\ldots ,A^{k-1}b\right\}\right\}}}
12791:
6070:
5781:
14586:
14557:
11190:
10025:
9621:
2359:
1713:
954:
827:
14179:
The conjugate gradient method with a trivial modification is extendable to solving, given complex-valued matrix A and vector b, the system of linear equations
9878:
6054:{\displaystyle \beta _{k}=-{\frac {\mathbf {r} _{k+1}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}}{\mathbf {p} _{k}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{k}}}}
14079:
13815:
methods is a bit cheaper compared to that for the conjugate gradient methods. However, the latter converge faster, unless a (highly) variable and/or non-SPD
12707:
11784:
14591:
9513:. If exact arithmetic were to be used in this example instead of limited-precision, then the exact solution would theoretically have been reached after
15574:
11979:
11683:
5433:
5062:
15502:
Gérard
Meurant: "Detection and correction of silent errors in the conjugate gradient algorithm", Numerical Algorithms, vol.92 (2023), pp.869-891. url=
11607:
11368:
4932:
1675:{\displaystyle \alpha _{k}={\frac {\langle \mathbf {p} _{k},\mathbf {b} \rangle }{\langle \mathbf {p} _{k},\mathbf {p} _{k}\rangle _{\mathbf {A} }}}.}
14882:
14148:) and so the rate of convergence of CGNR may be slow and the quality of the approximate solution may be sensitive to roundoff errors. Finding a good
2604:
2093:
14125:
explicitly in memory but only to perform the matrix–vector and transpose matrix–vector multiplications. Therefore, CGNR is particularly useful when
13838:
12346:
15395:
7564:{\displaystyle \mathbf {x} ={\begin{bmatrix}{\frac {1}{11}}\\\\{\frac {7}{11}}\end{bmatrix}}\approx {\begin{bmatrix}0.0909\\\\0.6364\end{bmatrix}}}
3477:
854:
10809:{\displaystyle k={\tfrac {1}{2}}{\sqrt {\kappa (\mathbf {A} )}}\log \left(\left\|\mathbf {e} _{0}\right\|_{\mathbf {A} }\varepsilon ^{-1}\right)}
13091:
653:
Two vectors are conjugate if and only if they are orthogonal with respect to this inner product. Being conjugate is a symmetric relation: if
410:
5362:
5288:
5140:
15695:
13662:
13152:
10106:
7402:
353:
13642:{\displaystyle \beta _{k}:={\frac {\mathbf {r} _{k+1}^{\mathsf {T}}\mathbf {z} _{k+1}}{\mathbf {r} _{k}^{\mathsf {T}}\mathbf {z} _{k}}}}
11968:{\displaystyle \beta _{k}:={\frac {\mathbf {r} _{k+1}^{\mathsf {T}}\mathbf {z} _{k+1}}{\mathbf {r} _{k}^{\mathsf {T}}\mathbf {z} _{k}}}}
1808:. So, we want to regard the conjugate gradient method as an iterative method. This also allows us to approximately solve systems where
11596:{\displaystyle \alpha _{k}:={\frac {\mathbf {r} _{k}^{\mathsf {T}}\mathbf {z} _{k}}{\mathbf {p} _{k}^{\mathsf {T}}\mathbf {Ap} _{k}}}}
11312:
5007:
3343:
2254:
15720:
13930:
9730:
747:
2025:
13652:
may dramatically improve the convergence in this case. This version of the preconditioned conjugate gradient method can be called
11071:
14758:
14242:
The advantages and disadvantages of the conjugate gradient methods are summarized in the lecture notes by
Nemirovsky and BenTal.
14706:
In words, during the CG process, the error grows exponentially, until it suddenly becomes zero as the unique solution is found.
12892:
12658:
15715:
15567:
12211:{\displaystyle \mathbf {E} ^{-1}\mathbf {A} (\mathbf {E} ^{-1})^{\mathsf {T}}\mathbf {\hat {x}} =\mathbf {E} ^{-1}\mathbf {b} }
11194:
14733:
12578:
15518:
15493:
15464:
15445:
15426:
15112:
15052:
15027:
14927:
11124:
is assumed in the convergence theorem, but the convergence bound is commonly valid in practice as theoretically explained by
9790:
14182:
12938:
12300:{\displaystyle \mathbf {EE} ^{\mathsf {T}}=\mathbf {M} ,\qquad \mathbf {\hat {x}} =\mathbf {E} ^{\mathsf {T}}\mathbf {x} .}
3982:
135:
15284:
Knyazev, Andrew V.; Lashuk, Ilya (2008). "Steepest
Descent and Conjugate Gradient Methods with Variable Preconditioning".
14748:
14133:
since these operations are usually extremely efficient. However the downside of forming the normal equations is that the
13737:
13531:
13384:
4743:
115:
11425:
10033:
15741:
15674:
15560:
14859:
14722:
12463:
15535:
1242:
15705:
13302:
13061:
12413:
12113:
The above formulation is equivalent to applying the regular conjugate gradient method to the preconditioned system
11250:
1868:
we search for the solution and in each iteration we need a metric to tell us whether we are closer to the solution
9691:
15632:
15159:
Golub, Gene H.; Ye, Qiang (1999). "Inexact
Preconditioned Conjugate Gradient Method with Inner-Outer Iteration".
11140:
13383:
practical tests leads to a significant slow down of the convergence of the algorithm presented above. Using the
13251:
13220:
12624:
1209:
1021:
4885:
4784:
8310:
This result completes the first iteration, the result being an "improved" approximate solution to the system,
15545:
13659:
The implementation of the flexible version requires storing an extra vector. For a fixed SPD preconditioner,
2241:
2225:
81:
15664:
13744:
in this case, in particular, it does not converge slower than the locally optimal steepest descent method.
10871:
10670:
10635:
9836:
6833:
6241:{\displaystyle \mathbf {A} \mathbf {p} _{k}={\frac {1}{\alpha _{k}}}(\mathbf {r} _{k}-\mathbf {r} _{k+1}),}
65:
15540:
14716:
111:
6475:
5871:
5530:
4852:
3758:
76:
systems that are too large to be handled by a direct implementation or other direct methods such as the
15618:
12499:
12314:
11218:
10842:
5229:
127:
61:
14993:
10077:
9544:
6510:
5906:
5565:
4891:
4790:
4757:
4039:
3704:
3653:
3624:
3591:
3562:
3448:
3397:
3314:
3285:
2891:
2406:
2364:
2315:
1782:
1753:
992:
963:
318:
15746:
15669:
14728:
88:
15583:
15210:
15173:
14219:
13755:
10819:
9529:
The conjugate gradient method can theoretically be viewed as a direct method, as in the absence of
5204:
mathematically equivalent. The former is used in the algorithm to avoid an extra multiplication by
2434:
253:
14738:
13355:
13333:
13311:
13282:
13198:
13067:
13041:
13019:
12997:
12975:
12868:
12554:
12531:
12477:
11287:
5359:
A norm of the residual is typically used for stopping criteria. The norm of the explicit residual
5207:
4819:
4017:
3733:
3682:
3426:
832:
722:
700:
678:
656:
460:
385:
296:
266:
221:
173:
14014:
13987:
13822:
11065:
5261:
2961:
2920:
1718:
15019:
15013:
9658:
9626:
9577:
15353:
15205:
15168:
957:
195:
77:
14787:
13752:
In both the original and the preconditioned conjugate gradient methods one only needs to set
12061:
9979:
15628:
14945:"Universality for the Conjugate Gradient and MINRES Algorithms on Sample Covariance Matrices"
10902:
3536:
913:
15018:(2nd ed.). Philadelphia, Pa.: Society for Industrial and Applied Mathematics. pp.
14253:
11470:
7396:
we will perform two steps of the conjugate gradient method beginning with the initial guess
5426:
provides a guaranteed level of accuracy both in exact arithmetic and in the presence of the
15623:
12856:{\displaystyle \mathbf {z} =(\mathbf {L} ^{-1})^{\mathsf {T}}\mathbf {L} ^{-1}\mathbf {r} }
10664:
9541:, the conjugate gradient method monotonically (in the energy norm) improves approximations
6140:{\displaystyle \mathbf {r} _{k+1}=\mathbf {r} _{k}-\alpha _{k}\mathbf {A} \mathbf {p} _{k}}
5854:{\displaystyle \mathbf {r} _{k+1}=\mathbf {p} _{k+1}-\mathbf {\beta } _{k}\mathbf {p} _{k}}
1779:
carefully, then we may not need all of them to obtain a good approximation to the solution
8:
15602:
14227:
14223:
480:
92:
14944:
9966:{\displaystyle \Pi _{k}^{*}:=\left\lbrace \ p\in \Pi _{k}\ :\ p(0)=1\ \right\rbrace \,,}
15482:
15476:
15415:
15335:
15311:
15293:
15266:
15246:
14956:
14542:
13736:
The mathematical explanation of the better convergence behavior of the method with the
11306:, a preconditioned conjugate gradient method can be used. It takes the following form:
11175:
10010:
9606:
2344:
1887:
1698:
939:
812:
57:
14562:
15514:
15489:
15460:
15441:
15422:
15389:
15108:
15058:
15048:
15023:
14974:
14923:
14886:
14855:
13924:
11172:, but may be super-linear, depending on a distribution of the spectrum of the matrix
363:
359:
15315:
15270:
15235:"Nonsymmetric Preconditioning for Conjugate Gradient and Steepest Descent Methods 1"
14156:
12777:{\displaystyle \mathbf {M} ^{-1}=(\mathbf {L} ^{-1})^{\mathsf {T}}\mathbf {L} ^{-1}}
11855:{\displaystyle \mathrm {Solve} \ \mathbf {M} \mathbf {z} _{k+1}:=\mathbf {r} _{k+1}}
15638:
15303:
15256:
15215:
15178:
15141:
15079:
Templates for the
Solution of Linear Systems: Building Blocks for Iterative Methods
14966:
14915:
14827:
14802:
14168:
14134:
14055:
13832:
13812:
13790:
10699:
9572:
9538:
2391:
2213:. The other vectors in the basis will be conjugate to the gradient, hence the name
241:
69:
24:
19:
15679:
15331:
14779:
14753:
14743:
13828:
13730:
12050:{\displaystyle \mathbf {p} _{k+1}:=\mathbf {z} _{k+1}+\beta _{k}\mathbf {p} _{k}}
11751:{\displaystyle \mathbf {r} _{k+1}:=\mathbf {r} _{k}-\alpha _{k}\mathbf {Ap} _{k}}
11212:
11121:
10004:
9685:
9530:
5509:
5501:{\displaystyle \mathbf {r} _{k+1}:=\mathbf {r} _{k}-\alpha _{k}\mathbf {Ap} _{k}}
5427:
5353:
5130:{\displaystyle \mathbf {r} _{k+1}:=\mathbf {r} _{k}-\alpha _{k}\mathbf {Ap} _{k}}
4918:
3618:
96:
15261:
15234:
15077:
11672:{\displaystyle \mathbf {x} _{k+1}:=\mathbf {x} _{k}+\alpha _{k}\mathbf {p} _{k}}
11414:{\displaystyle {\textrm {Solve:}}\mathbf {M} \mathbf {z} _{0}:=\mathbf {r} _{0}}
11060:
This limit shows a faster convergence rate compared to the iterative methods of
4997:{\displaystyle \mathbf {x} _{k+1}:=\mathbf {x} _{k}+\alpha _{k}\mathbf {p} _{k}}
15597:
14907:
14847:
14783:
14149:
13816:
12459:
11206:
11125:
2669:{\displaystyle \mathbf {x} _{k+1}=\mathbf {x} _{k}+\alpha _{k}\mathbf {p} _{k}}
2142:{\displaystyle \nabla f(\mathbf {x} )=\mathbf {A} \mathbf {x} -\mathbf {b} \,.}
2016:
100:
15219:
15182:
13823:
Conjugate gradient method as optimal feedback controller for double integrator
12403:{\displaystyle \mathbf {E} ^{-1}\mathbf {A} (\mathbf {E} ^{-1})^{\mathsf {T}}}
121:
15735:
15643:
15503:
15062:
14978:
14130:
11215:
is necessary to ensure fast convergence of the conjugate gradient method. If
11061:
104:
73:
15371:
15129:
14919:
14832:
9534:
conjugate, due to a degenerative nature of generating the Krylov subspaces.
5430:, where convergence naturally stagnates. In contrast, the implicit residual
3526:{\displaystyle \mathbf {p} _{i}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{j}=0}
903:{\displaystyle \mathbf {p} _{i}^{\mathsf {T}}\mathbf {A} \mathbf {p} _{j}=0}
15552:
14995:
An
Introduction to the Conjugate Gradient Method Without the Agonizing Pain
14588:. Solving it by conjugate gradient descent gives us rather bad convergence:
14174:
6817:
are conjugated and again that the residuals are orthogonal. This gives the
1877:(that is unknown to us). This metric comes from the fact that the solution
13141:{\displaystyle \mathbf {z} =(\mathbf {L} ^{-1})^{\mathsf {T}}\mathbf {a} }
3650:
form the orthogonal basis with respect to the standard inner product, and
14807:
13786:
5285:. The latter may be more accurate, substituting the explicit calculation
290:
45:
13377:
12474:
It is importart to keep in mind that we don't want to invert the matrix
4069:. It is a different formulation of the exact procedure described above.
15370:
Pennington, Fabian
Pedregosa, Courtney Paquette, Tom Trogdon, Jeffrey.
6847:
Return the solution to `A * x = b` using the conjugate gradient method.
3679:
form the orthogonal basis with respect to the inner product induced by
367:
15307:
14970:
5508:
is known to keep getting smaller in amplitude well below the level of
87:
The conjugate gradient method can also be used to solve unconstrained
15659:
15298:
15233:
Bouwmeester, Henricus; Dougherty, Andrew; Knyazev, Andrew V. (2015).
14231:
14215:
447:{\displaystyle \mathbf {u} ^{\mathsf {T}}\mathbf {A} \mathbf {v} =0.}
53:
15145:
8564:
that will eventually be used to determine the next search direction
5512:
and thus cannot be used to determine the stagnation of convergence.
5419:{\displaystyle \mathbf {r} _{k+1}:=\mathbf {b} -\mathbf {Ax} _{k+1}}
5356:
accumulation, and is thus recommended for an occasional evaluation.
5345:{\displaystyle \mathbf {r} _{k+1}:=\mathbf {b} -\mathbf {Ax} _{k+1}}
5197:{\displaystyle \mathbf {r} _{k+1}:=\mathbf {b} -\mathbf {Ax} _{k+1}}
15339:
14961:
14699:{\displaystyle \|b-Wx_{k}\|^{2}=(1/t)^{k},\quad \|b-Wx_{n}\|^{2}=0}
13715:{\displaystyle \mathbf {r} _{k+1}^{\mathsf {T}}\mathbf {z} _{k}=0,}
13186:{\displaystyle \mathbf {a} =\mathbf {L} ^{\mathsf {T}}\mathbf {z} }
10155:{\displaystyle \mathbf {e} _{k}:=\mathbf {x} _{k}-\mathbf {x} _{*}}
7788:
Since this is the first iteration, we will use the residual vector
7450:{\displaystyle \mathbf {x} _{0}={\begin{bmatrix}2\\1\end{bmatrix}}}
15251:
13831:. In this approach, the conjugate gradient method falls out as an
13747:
2087:)=0) solves the initial problem follows from its first derivative
479:
is symmetric and positive-definite, the left-hand side defines an
315:
is known as well. We denote the unique solution of this system by
4036:
is a real, symmetric, positive-definite matrix. The input vector
11357:{\displaystyle \mathbf {r} _{0}:=\mathbf {b} -\mathbf {Ax} _{0}}
5052:{\displaystyle \mathbf {r} _{k}:=\mathbf {b} -\mathbf {Ax} _{k}}
3387:{\displaystyle \mathbf {r} _{i}^{\mathsf {T}}\mathbf {r} _{j}=0}
2302:{\displaystyle \mathbf {r} _{k}=\mathbf {b} -\mathbf {Ax} _{k}.}
15710:
15700:
11200:
9780:{\displaystyle \mathbf {M} ^{-1}(\mathbf {Ax} -\mathbf {b} )=0}
6831:
4735:
This is the most commonly used algorithm. The same formula for
14155:
Several algorithms have been proposed (e.g., CGLS, LSQR). The
13916:{\displaystyle u=k(x,v):=-\gamma _{a}\nabla f(x)-\gamma _{b}v}
799:{\displaystyle P=\{\mathbf {p} _{1},\dots ,\mathbf {p} _{n}\}}
15334:, "An Optimal Control Theory for Accelerated Optimization,"
15107:(4th ed.). Johns Hopkins University Press. sec. 11.5.2.
14046:
The conjugate gradient method can be applied to an arbitrary
14041:
2069:{\displaystyle \mathbf {H} (f(\mathbf {x} ))=\mathbf {A} \,,}
1812:
is so large that the direct method would take too much time.
114:
provides a generalization to non-symmetric matrices. Various
14852:
Geschichten der Informatik. Visionen, Paradigmen, Leitmotive
14159:
algorithm purportedly has the best numerical stability when
11110:{\displaystyle \approx 1-{\frac {2}{\kappa (\mathbf {A} )}}}
9499:, is a "better" approximation to the system's solution than
80:. Large sparse systems often arise when numerically solving
15455:
Golub, Gene H.; Van Loan, Charles F. (2013). "Chapter 11".
15232:
14788:"Methods of Conjugate Gradients for Solving Linear Systems"
122:
Description of the problem addressed by conjugate gradients
68:. The conjugate gradient method is often implemented as an
12927:{\displaystyle \mathbf {a} =\mathbf {L} ^{-1}\mathbf {r} }
12693:{\displaystyle \mathbf {z} =\mathbf {M} ^{-1}\mathbf {r} }
8317:. We may now move on and compute the next residual vector
6536:
are orthogonal by design. The denominator is rewritten as
15511:
Error Norm Estimation in the Conjugate Gradient Algorithm
15130:"Block Preconditioning for the Conjugate Gradient Method"
5059:, which both hold in exact arithmetic, make the formulas
3559:. This can be regarded that as the algorithm progresses,
1685:
This gives the following method for solving the equation
14175:
Conjugate gradient method for complex Hermitian matrices
13827:
The conjugate gradient method can also be derived using
12610:{\displaystyle \mathbf {M} =\mathbf {LL} ^{\mathsf {T}}}
7460:
in order to find an approximate solution to the system.
1715:
conjugate directions, and then compute the coefficients
15045:
Iterative solution of large sparse systems of equations
14795:
Journal of Research of the National Bureau of Standards
14778:
12469:
10162:. Now, the rate of convergence can be approximated as
3282:
However, a closer analysis of the algorithm shows that
2888:
where the last equality follows from the definition of
2015:
The existence of a unique minimizer is apparent as its
15351:
14565:
14494:
14306:
14256:
14226:. The trivial modification is simply substituting the
13729:
are equivalent in exact arithmetic, i.e., without the
11143:
10719:
10074:
be the iterative approximations of the exact solution
9826:{\displaystyle \kappa (\mathbf {M} ^{-1}\mathbf {A} )}
9455:
9423:
9391:
9261:
9225:
9198:
9167:
9140:
8968:
8933:
8898:
8763:
8733:
8702:
8675:
8555:
Our next step in the process is to compute the scalar
8515:
8480:
8444:
8399:
8273:
8238:
8199:
8060:
8024:
7994:
7960:
7930:
7730:
7701:
7665:
7636:
7537:
7491:
7426:
7359:
7316:
7280:
4454:
4128:
1917:
14594:
14545:
14294:
14207:{\displaystyle \mathbf {A} \mathbf {x} =\mathbf {b} }
14185:
14152:
is often an important part of using the CGNR method.
14118:
As an iterative method, it is not necessary to form
14017:
13990:
13933:
13841:
13758:
13665:
13543:
13396:
13378:
The flexible preconditioned conjugate gradient method
13358:
13336:
13314:
13285:
13254:
13223:
13201:
13155:
13094:
13070:
13044:
13022:
13000:
12978:
12963:{\displaystyle \mathbf {r} =\mathbf {L} \mathbf {a} }
12941:
12895:
12871:
12794:
12710:
12661:
12627:
12581:
12557:
12534:
12502:
12480:
12416:
12349:
12317:
12230:
12122:
12064:
11982:
11869:
11787:
11686:
11610:
11506:
11473:
11428:
11371:
11315:
11290:
11253:
11221:
11178:
11074:
10927:
10905:
10874:
10845:
10822:
10711:
10673:
10638:
10171:
10109:
10080:
10036:
10013:
9982:
9881:
9839:
9793:
9733:
9694:
9661:
9629:
9609:
9580:
9547:
9330:
9042:
8837:
8580:
8333:
8138:
7832:
7615:
7477:
7405:
7261:
6545:
6513:
6478:
6266:
6158:
6073:
5948:
5909:
5874:
5784:
5600:
5568:
5533:
5436:
5365:
5291:
5264:
5232:
5210:
5143:
5065:
5010:
4935:
4894:
4855:
4822:
4793:
4760:
4078:
4042:
4020:
4007:{\displaystyle \mathbf {A} \mathbf {x} =\mathbf {b} }
3985:
3796:
3761:
3736:
3707:
3685:
3656:
3627:
3594:
3565:
3539:
3480:
3451:
3429:
3400:
3346:
3317:
3288:
2993:
2964:
2944:
can be derived if one substitutes the expression for
2923:
2894:
2688:
2607:
2446:
2409:
2367:
2347:
2318:
2257:
2096:
2028:
2019:
of second derivatives is symmetric positive-definite
1898:
1785:
1756:
1721:
1701:
1579:
1284:
1245:
1212:
1060:
1024:
995:
966:
942:
916:
857:
835:
815:
750:
725:
703:
681:
659:
491:
463:
413:
388:
321:
299:
269:
224:
198:
176:
160:{\displaystyle \mathbf {A} \mathbf {x} =\mathbf {b} }
138:
15372:"Random Matrix Theory and Machine Learning Tutorial"
15196:
Notay, Yvan (2000). "Flexible Conjugate Gradients".
15134:
SIAM Journal on Scientific and Statistical Computing
7574:Our first step is to calculate the residual vector
15481:
15475:
15414:
15127:
15085:(2nd ed.). Philadelphia, PA: SIAM. p. 13
15075:
14698:
14580:
14551:
14531:
14280:
14206:
14030:
14003:
13976:
13915:
13777:
13714:
13641:
13519:
13366:
13344:
13322:
13293:
13271:
13240:
13209:
13185:
13140:
13078:
13052:
13030:
13008:
12986:
12962:
12926:
12879:
12855:
12776:
12692:
12646:
12609:
12565:
12542:
12520:
12488:
12439:
12402:
12335:
12299:
12210:
12083:
12049:
11967:
11854:
11750:
11671:
11595:
11486:
11459:{\displaystyle \mathbf {p} _{0}:=\mathbf {z} _{0}}
11458:
11413:
11356:
11298:
11276:
11239:
11184:
11164:
11109:
11049:
10911:
10891:
10857:
10831:
10808:
10690:
10655:
10621:
10154:
10095:
10066:
10019:
9995:
9965:
9856:
9825:
9779:
9719:
9676:
9644:
9615:
9595:
9562:
9481:
9296:
8997:
8801:
8544:
8299:
8111:
7777:
7563:
7449:
7385:
6797:
6528:
6499:
6461:
6240:
6139:
6053:
5924:
5895:
5853:
5767:
5583:
5554:
5500:
5418:
5344:
5277:
5250:
5218:
5196:
5129:
5051:
4996:
4909:
4876:
4841:
4808:
4775:
4724:
4057:
4028:
4006:
3971:
3782:
3744:
3722:
3693:
3671:
3642:
3609:
3580:
3551:
3525:
3466:
3437:
3415:
3386:
3332:
3303:
3265:
2977:
2936:
2909:
2877:
2668:
2587:
2424:
2382:
2353:
2333:
2301:
2141:
2068:
2004:
1800:
1771:
1734:
1707:
1674:
1559:
1267:
1231:
1195:
1043:
1010:
981:
948:
928:
902:
843:
821:
798:
733:
711:
689:
667:
642:
471:
446:
396:
336:
307:
277:
232:
210:
184:
159:
14559:is invertible, there exists a unique solution to
13785:in order to make them locally optimal, using the
12266:
12179:
10067:{\displaystyle \left(\mathbf {x} _{k}\right)_{k}}
7213:# Update squared residual norm for next iteration
5352:for the implicit one by the recursion subject to
15733:
15459:(4th ed.). Johns Hopkins University Press.
15286:SIAM Journal on Matrix Analysis and Applications
15005:
14943:Paquette, Elliot; Trogdon, Thomas (March 2023).
10325:
10296:
10213:
9688:is commonly used to replace the original system
5515:
4924:
2228:provided by this initial step of the algorithm.
347:
118:seek minima of nonlinear optimization problems.
14942:
14826:(PhD thesis). North Carolina State University.
14237:
13977:{\displaystyle {\dot {x}}=v,\quad {\dot {v}}=u}
13748:Vs. the locally optimal steepest descent method
2403:. Here, however, we insist that the directions
1833:(we can assume without loss of generality that
15128:Concus, P.; Golub, G. H.; Meurant, G. (1985).
14949:Communications on Pure and Applied Mathematics
13308:Using this method, there is no need to invert
7588:. This residual is computed from the formula
4148: is sufficiently small, then return
2394:method would require to move in the direction
1886:is also the unique minimizer of the following
1268:{\displaystyle \mathbf {p} _{k}^{\mathsf {T}}}
36:is the size of the matrix of the system (here
15568:
15454:
15283:
15103:Golub, Gene H.; Van Loan, Charles F. (2013).
15102:
12440:{\displaystyle \mathbf {M} ^{-1}\mathbf {A} }
11277:{\displaystyle \mathbf {M} ^{-1}\mathbf {A} }
15582:
15508:
15413:Atkinson, Kendell A. (1988). "Section 8.9".
15394:: CS1 maint: multiple names: authors list (
14912:Iterative Methods for Solving Linear Systems
14854:(in German). Berlin: Springer. p. 185.
14681:
14658:
14618:
14595:
14214:for the complex-valued vector x, where A is
11201:The preconditioned conjugate gradient method
11165:{\textstyle {\sqrt {\kappa (\mathbf {A} )}}}
9720:{\displaystyle \mathbf {Ax} -\mathbf {b} =0}
3979:The algorithm is detailed below for solving
2152:This suggests taking the first basis vector
1655:
1624:
1619:
1596:
793:
757:
634:
613:
607:
577:
571:
550:
536:
519:
15484:Iterative methods for sparse linear systems
15438:Nonlinear Programming: Analysis and Methods
15015:Iterative methods for sparse linear systems
13301:is easy and computationally cheap by using
13060:is easy and computationally cheap by using
12528:for use it in the process, since inverting
10816:iterations suffices to reduce the error to
9314:using the same method as that used to find
8821:, we can compute the next search direction
4480: is sufficiently small, then exit loop
829:mutually conjugate vectors with respect to
354:Derivation of the conjugate gradient method
15575:
15561:
15504:https://doi.org/10.1007/s11075-022-01380-1
14836:– via NASA Technical Reports Server.
14088:conjugate gradient on the normal equations
14042:Conjugate gradient on the normal equations
13272:{\displaystyle \mathbf {L} ^{\mathsf {T}}}
13241:{\displaystyle \mathbf {L} ^{\mathsf {T}}}
12647:{\displaystyle \mathbf {Mz} =\mathbf {r} }
4065:can be an approximate initial solution or
3276:
1232:{\displaystyle \mathbf {Ax} =\mathbf {b} }
1044:{\displaystyle \mathbf {Ax} =\mathbf {b} }
15297:
15260:
15250:
15209:
15172:
15042:
14960:
14906:
14831:
14806:
14245:
13793:methods. With this substitution, vectors
12080:
11483:
11043:
10611:
10323:
9959:
9524:
3840:
3255:
2135:
2062:
1998:
1745:
969:
15412:
14991:
14821:
11131:
18:
15706:Basic Linear Algebra Subprograms (BLAS)
15354:"Optimization III: Convex Optimization"
15158:
15047:(2nd ed.). Switzerland: Springer.
14846:
13805:, so there is no need to store vectors
13352:explicitly at all, and we still obtain
9024:using the same method as that used for
6970:# Compute initial squared residual norm
2437:. This gives the following expression:
15734:
15435:
15369:
14881:
13685:
13618:
13579:
13496:
13432:
13263:
13232:
13172:
13127:
12827:
12753:
12601:
12394:
12283:
12242:
12167:
11944:
11905:
11569:
11536:
11195:double-precision floating-point format
9867:
9105:
9072:
8643:
8610:
7895:
7862:
7468:For reference, the exact solution is
6777:
6699:
6643:
6559:
6435:
6348:
6286:
6025:
5987:
5741:
5708:
5663:
5630:
4563:
4524:
4292:
4259:
3755:That is, if the CG method starts with
3494:
3360:
3229:
3179:
2846:
2813:
2768:
2718:
2547:
2509:
2159:to be the negative of the gradient of
1963:
1936:
1395:
1325:
1298:
1259:
871:
596:
500:
422:
15556:
15509:Meurant, Gerard; Tichy, Petr (2024).
15421:(2nd ed.). John Wiley and Sons.
15417:An introduction to numerical analysis
15327:
15325:
15195:
10892:{\displaystyle \kappa (\mathbf {A} )}
10691:{\displaystyle \kappa (\mathbf {A} )}
10656:{\displaystyle \sigma (\mathbf {A} )}
9857:{\displaystyle \kappa (\mathbf {A} )}
5591:. The denominator is simplified from
3730:can be regarded as the projection of
15473:
15198:SIAM Journal on Scientific Computing
15161:SIAM Journal on Scientific Computing
15011:
14902:
14900:
14898:
14877:
14875:
14873:
14871:
12573:, and the preconditioner matrix is:
12470:Using the preconditioner in practice
7239:
6949:# Initialize search direction vector
4742:is also used in the Fletcher–Reeves
116:nonlinear conjugate gradient methods
15345:
14759:Sparse matrix–vector multiplication
14749:Nonlinear conjugate gradient method
11284:has a better condition number than
11247:is symmetric positive-definite and
7811:will change in further iterations.
4744:nonlinear conjugate gradient method
1750:If we choose the conjugate vectors
23:A comparison of the convergence of
16:Mathematical optimization algorithm
13:
15406:
15322:
13882:
12865:Let's take an intermediary vector
11801:
11798:
11795:
11792:
11789:
10906:
10307:
10224:
9984:
9915:
9883:
9872:Define a subset of polynomials as
6821:in the algorithm after cancelling
6500:{\displaystyle \mathbf {r} _{k+1}}
5896:{\displaystyle \mathbf {p} _{k+1}}
5555:{\displaystyle \mathbf {r} _{k+1}}
4877:{\displaystyle \mathbf {x} _{k+1}}
3875:
3827:
3824:
3821:
3818:
3815:
3812:
3783:{\displaystyle \mathbf {x} _{0}=0}
2958:and minimizing it with respect to
2097:
989:, and we may express the solution
14:
15758:
15528:
14895:
14868:
14723:Conjugate gradient squared method
13811:. Thus, every iteration of these
13279:is upper triangular, solving for
13038:is lower triangular, solving for
12521:{\displaystyle \mathbf {M} ^{-1}}
12464:incomplete Cholesky factorization
12336:{\displaystyle \mathbf {M} ^{-1}}
11240:{\displaystyle \mathbf {M} ^{-1}}
10858:{\displaystyle \varepsilon >0}
6808:using that the search directions
5251:{\displaystyle \mathbf {Ap} _{k}}
2188:. Starting with an initial guess
374:We say that two non-zero vectors
107:, and extensively researched it.
15536:"Conjugate gradients, method of"
15376:random-matrix-learning.github.io
14739:Iterative method: Linear systems
14200:
14192:
14187:
13693:
13668:
13626:
13607:
13587:
13562:
13504:
13485:
13466:
13445:
13415:
13360:
13338:
13316:
13287:
13257:
13226:
13203:
13179:
13166:
13157:
13134:
13108:
13096:
13072:
13046:
13024:
13002:
12980:
12956:
12951:
12943:
12920:
12906:
12897:
12873:
12849:
12835:
12808:
12796:
12761:
12734:
12713:
12686:
12672:
12663:
12640:
12632:
12629:
12595:
12592:
12583:
12559:
12536:
12505:
12482:
12433:
12419:
12375:
12366:
12352:
12320:
12290:
12277:
12263:
12252:
12236:
12233:
12204:
12190:
12176:
12148:
12139:
12125:
12037:
12006:
11985:
11952:
11933:
11913:
11888:
11836:
11815:
11809:
11738:
11735:
11710:
11689:
11659:
11634:
11613:
11580:
11577:
11558:
11544:
11525:
11446:
11431:
11401:
11386:
11380:
11344:
11341:
11332:
11318:
11292:
11270:
11256:
11224:
11153:
11097:
11030:
11006:
10966:
10940:
10882:
10782:
10766:
10739:
10681:
10646:
10605:
10589:
10565:
10515:
10499:
10461:
10435:
10399:
10383:
10342:
10279:
10262:
10253:
10199:
10183:
10142:
10127:
10112:
10096:{\displaystyle \mathbf {x} _{*}}
10083:
10044:
9847:
9816:
9802:
9764:
9756:
9753:
9736:
9707:
9699:
9696:
9652:is, the slower the improvement.
9563:{\displaystyle \mathbf {x} _{k}}
9550:
9521:being the order of the system).
9373:
9348:
9333:
9116:
9113:
9094:
9080:
9061:
8880:
8855:
8840:
8651:
8632:
8618:
8599:
8381:
8375:
8351:
8336:
8181:
8156:
8141:
7906:
7903:
7884:
7870:
7851:
7795:as our initial search direction
7762:
7618:
7479:
7408:
7268:
7263:
7168:# Update search direction vector
6785:
6766:
6725:
6710:
6688:
6656:
6650:
6621:
6590:
6572:
6566:
6548:
6529:{\displaystyle \mathbf {r} _{k}}
6516:
6481:
6443:
6418:
6374:
6359:
6331:
6299:
6293:
6269:
6216:
6201:
6166:
6160:
6127:
6121:
6097:
6076:
6038:
6032:
6014:
6000:
5994:
5970:
5925:{\displaystyle \mathbf {p} _{k}}
5912:
5877:
5841:
5808:
5787:
5752:
5749:
5730:
5716:
5697:
5676:
5670:
5652:
5638:
5619:
5584:{\displaystyle \mathbf {r} _{k}}
5571:
5536:
5488:
5485:
5460:
5439:
5400:
5397:
5388:
5368:
5326:
5323:
5314:
5294:
5258:is already computed to evaluate
5238:
5235:
5212:
5178:
5175:
5166:
5146:
5117:
5114:
5089:
5068:
5039:
5036:
5027:
5013:
4984:
4959:
4938:
4910:{\displaystyle \mathbf {x} _{k}}
4897:
4858:
4809:{\displaystyle \mathbf {x} _{0}}
4796:
4776:{\displaystyle \mathbf {x} _{1}}
4763:
4697:
4644:
4613:
4592:
4571:
4552:
4532:
4507:
4462:
4437:
4434:
4409:
4388:
4370:
4345:
4324:
4303:
4300:
4281:
4267:
4248:
4190:
4175:
4153:
4136:
4112:
4109:
4100:
4086:
4058:{\displaystyle \mathbf {x} _{0}}
4045:
4022:
4000:
3992:
3987:
3764:
3738:
3723:{\displaystyle \mathbf {x} _{k}}
3710:
3687:
3672:{\displaystyle \mathbf {p} _{i}}
3659:
3643:{\displaystyle \mathbf {r} _{i}}
3630:
3610:{\displaystyle \mathbf {r} _{i}}
3597:
3581:{\displaystyle \mathbf {p} _{i}}
3568:
3507:
3501:
3483:
3467:{\displaystyle \mathbf {p} _{j}}
3454:
3431:
3416:{\displaystyle \mathbf {p} _{i}}
3403:
3368:
3349:
3333:{\displaystyle \mathbf {r} _{j}}
3320:
3304:{\displaystyle \mathbf {r} _{i}}
3291:
3242:
3236:
3218:
3201:
3198:
3189:
3168:
3065:
3040:
3006:
2910:{\displaystyle \mathbf {r} _{k}}
2897:
2859:
2853:
2835:
2821:
2802:
2781:
2775:
2757:
2740:
2737:
2728:
2707:
2656:
2631:
2610:
2575:
2560:
2554:
2536:
2522:
2516:
2498:
2464:
2449:
2425:{\displaystyle \mathbf {p} _{k}}
2412:
2383:{\displaystyle \mathbf {x} _{k}}
2370:
2334:{\displaystyle \mathbf {r} _{k}}
2321:
2286:
2283:
2274:
2260:
2131:
2123:
2118:
2107:
2058:
2044:
2030:
1988:
1979:
1970:
1957:
1948:
1943:
1930:
1906:
1846:, otherwise consider the system
1815:We denote the initial guess for
1801:{\displaystyle \mathbf {x} _{*}}
1788:
1772:{\displaystyle \mathbf {p} _{k}}
1759:
1660:
1644:
1629:
1615:
1601:
1551:
1534:
1519:
1492:
1475:
1460:
1408:
1402:
1384:
1338:
1332:
1314:
1305:
1287:
1248:
1225:
1217:
1214:
1180:
1174:
1129:
1123:
1109:
1063:
1037:
1029:
1026:
1011:{\displaystyle \mathbf {x} _{*}}
998:
982:{\displaystyle \mathbb {R} ^{n}}
884:
878:
860:
837:
783:
762:
727:
705:
683:
661:
630:
625:
617:
603:
590:
581:
567:
559:
554:
541:
531:
523:
512:
507:
494:
465:
434:
429:
416:
390:
337:{\displaystyle \mathbf {x} _{*}}
324:
301:
271:
263:> 0 for all non-zero vectors
226:
178:
153:
145:
140:
15363:
15352:Nemirovsky and Ben-Tal (2023).
15277:
15226:
15189:
15152:
15121:
15096:
15069:
14657:
14482:
14222:, and the symbol ' denotes the
13955:
13799:are always the same as vectors
12259:
11022:
11016:
10868:Note, the important limit when
4659:
4589:
4488:
4452:
4385:
4321:
4229:
3149:
3145:
2435:Gram-Schmidt orthonormalization
1977:
382:are conjugate (with respect to
95:. It is commonly attributed to
64:, namely those whose matrix is
15036:
14985:
14936:
14840:
14815:
14772:
14645:
14630:
14275:
14263:
13894:
13888:
13863:
13851:
13740:formula is that the method is
13122:
13103:
12822:
12803:
12748:
12729:
12458:An example of a commonly used
12389:
12370:
12162:
12143:
11157:
11149:
11101:
11093:
11034:
11026:
11010:
11002:
10970:
10962:
10944:
10936:
10886:
10878:
10776:
10761:
10743:
10735:
10685:
10677:
10650:
10642:
10599:
10584:
10569:
10561:
10509:
10494:
10465:
10457:
10439:
10431:
10393:
10378:
10369:
10365:
10359:
10352:
10346:
10338:
10273:
10257:
10249:
10242:
10193:
10178:
9942:
9936:
9851:
9843:
9820:
9797:
9768:
9749:
9671:
9665:
9639:
9633:
9590:
9584:
7606:, and in our case is equal to
6741:
6705:
6638:
6585:
6390:
6354:
6232:
6196:
3895:
3883:
3871:
3858:
3211:
3185:
3146:
3125:
3112:
3097:
3084:
3075:
3035:
3022:
3001:
2750:
2724:
2111:
2103:
2051:
2048:
2040:
2034:
1910:
1902:
1119:
82:partial differential equations
1:
14992:Shewchuk, Jonathan R (1994).
14765:
14038:are variable feedback gains.
13778:{\displaystyle \beta _{k}:=0}
10832:{\displaystyle 2\varepsilon }
5516:Computation of alpha and beta
4925:Explicit residual calculation
2079:and that the minimizer (use D
1206:Left-multiplying the problem
348:Derivation as a direct method
126:Suppose we want to solve the
15043:Hackbusch, W. (2016-06-21).
14888:Introduction to Optimization
14238:Advantages and disadvantages
13367:{\displaystyle \mathbf {z} }
13345:{\displaystyle \mathbf {L} }
13323:{\displaystyle \mathbf {M} }
13294:{\displaystyle \mathbf {z} }
13210:{\displaystyle \mathbf {a} }
13079:{\displaystyle \mathbf {a} }
13053:{\displaystyle \mathbf {a} }
13031:{\displaystyle \mathbf {L} }
13009:{\displaystyle \mathbf {L} }
12987:{\displaystyle \mathbf {r} }
12880:{\displaystyle \mathbf {a} }
12566:{\displaystyle \mathbf {L} }
12543:{\displaystyle \mathbf {M} }
12489:{\displaystyle \mathbf {M} }
11299:{\displaystyle \mathbf {A} }
6925:# Initialize residual vector
6844:conjugate_gradient!(A, b, x)
6834:Julia (programming language)
5219:{\displaystyle \mathbf {A} }
4842:{\displaystyle \beta _{k}=0}
4029:{\displaystyle \mathbf {A} }
3745:{\displaystyle \mathbf {x} }
3694:{\displaystyle \mathbf {A} }
3438:{\displaystyle \mathbf {A} }
2341:is the negative gradient of
844:{\displaystyle \mathbf {A} }
734:{\displaystyle \mathbf {u} }
712:{\displaystyle \mathbf {v} }
690:{\displaystyle \mathbf {v} }
668:{\displaystyle \mathbf {u} }
472:{\displaystyle \mathbf {A} }
397:{\displaystyle \mathbf {A} }
308:{\displaystyle \mathbf {b} }
278:{\displaystyle \mathbf {x} }
233:{\displaystyle \mathbf {A} }
185:{\displaystyle \mathbf {x} }
7:
15541:Encyclopedia of Mathematics
15474:Saad, Yousef (2003-04-01).
15262:10.1016/j.procs.2015.05.241
14734:Gaussian belief propagation
14717:Biconjugate gradient method
14709:
14064:and right-hand side vector
14031:{\displaystyle \gamma _{b}}
14004:{\displaystyle \gamma _{a}}
13833:optimal feedback controller
12496:explicitly in order to get
10103:, and define the errors as
7463:
7244:Consider the linear system
7030:# Iterate until convergence
5278:{\displaystyle \alpha _{k}}
4749:
2978:{\displaystyle \alpha _{k}}
2937:{\displaystyle \alpha _{k}}
1735:{\displaystyle \alpha _{k}}
112:biconjugate gradient method
103:, who programmed it on the
62:systems of linear equations
10:
15763:
15619:System of linear equations
14250:This example is from Let
14163:is ill-conditioned, i.e.,
13086:in the original equation:
12450:The preconditioner matrix
11204:
9677:{\displaystyle \kappa (A)}
9645:{\displaystyle \kappa (A)}
9596:{\displaystyle \kappa (A)}
9008:We now compute the scalar
7814:We now compute the scalar
7802:; the method of selecting
351:
128:system of linear equations
84:or optimization problems.
15688:
15670:Cache-oblivious algorithm
15652:
15611:
15590:
15436:Avriel, Mordecai (2003).
15239:Procedia Computer Science
15220:10.1137/S1064827599362314
15183:10.1137/S1064827597323415
14729:Conjugate residual method
14054:matrix by applying it to
12410:has the same spectrum as
9017:using our newly acquired
2215:conjugate gradient method
211:{\displaystyle n\times n}
50:conjugate gradient method
15742:Numerical linear algebra
15721:General purpose software
15584:Numerical linear algebra
14822:Straeter, T. A. (1971).
14220:positive-definite matrix
13925:double integrator system
12084:{\displaystyle k:=k+1\,}
9996:{\displaystyle \Pi _{k}}
6838:
3752:on the Krylov subspace.
1861:instead). Starting with
14920:10.1137/1.9781611970937
14281:{\textstyle t\in (0,1)}
12619:Then we have to solve:
12343:. It can be shown that
10912:{\displaystyle \infty }
8828:using the relationship
8812:Now, using this scalar
7823:using the relationship
3552:{\displaystyle i\neq j}
3277:The resulting algorithm
929:{\displaystyle i\neq j}
15488:(2nd ed.). SIAM.
14700:
14582:
14553:
14533:
14282:
14246:A pathological example
14208:
14032:
14005:
13978:
13917:
13829:optimal control theory
13779:
13716:
13643:
13521:
13368:
13346:
13324:
13295:
13273:
13242:
13211:
13187:
13142:
13080:
13064:. Then, we substitute
13054:
13032:
13010:
12988:
12964:
12928:
12881:
12857:
12778:
12694:
12648:
12611:
12567:
12544:
12522:
12490:
12441:
12404:
12337:
12301:
12212:
12085:
12051:
11969:
11856:
11772:is sufficiently small
11752:
11673:
11597:
11488:
11487:{\displaystyle k:=0\,}
11460:
11415:
11358:
11300:
11278:
11241:
11186:
11166:
11111:
11051:
10913:
10893:
10859:
10833:
10810:
10692:
10657:
10623:
10156:
10097:
10068:
10021:
9997:
9967:
9858:
9827:
9781:
9721:
9678:
9646:
9617:
9597:
9564:
9525:Convergence properties
9483:
9298:
8999:
8803:
8546:
8301:
8113:
7779:
7565:
7451:
7387:
6799:
6530:
6501:
6463:
6242:
6141:
6055:
5926:
5897:
5855:
5769:
5585:
5556:
5502:
5420:
5346:
5279:
5252:
5220:
5198:
5131:
5053:
4998:
4911:
4878:
4843:
4810:
4777:
4726:
4059:
4030:
4008:
3973:
3784:
3746:
3724:
3695:
3673:
3644:
3611:
3582:
3553:
3527:
3468:
3439:
3417:
3388:
3334:
3305:
3267:
2979:
2938:
2911:
2879:
2670:
2589:
2426:
2384:
2355:
2335:
2303:
2143:
2070:
2006:
1802:
1773:
1746:As an iterative method
1736:
1709:
1676:
1561:
1441:
1371:
1269:
1233:
1197:
1162:
1096:
1045:
1012:
983:
950:
930:
904:
845:
823:
800:
735:
713:
691:
669:
644:
473:
448:
398:
338:
309:
279:
234:
212:
186:
161:
78:Cholesky decomposition
41:
15716:Specialized libraries
15629:Matrix multiplication
15624:Matrix decompositions
15012:Saad, Yousef (2003).
14701:
14583:
14554:
14534:
14283:
14209:
14080:positive-semidefinite
14033:
14006:
13979:
13918:
13780:
13722:so both formulas for
13717:
13644:
13522:
13369:
13347:
13325:
13303:backward substitution
13296:
13274:
13243:
13212:
13188:
13143:
13081:
13055:
13033:
13011:
12989:
12965:
12929:
12882:
12858:
12779:
12695:
12649:
12612:
12568:
12545:
12523:
12491:
12442:
12405:
12338:
12302:
12213:
12086:
12052:
11970:
11857:
11753:
11674:
11598:
11489:
11461:
11416:
11359:
11301:
11279:
11242:
11187:
11167:
11132:Practical convergence
11112:
11052:
10914:
10894:
10860:
10834:
10811:
10693:
10658:
10624:
10157:
10098:
10069:
10022:
9998:
9968:
9859:
9828:
9782:
9722:
9679:
9647:
9618:
9603:of the system matrix
9598:
9565:
9484:
9299:
9000:
8804:
8547:
8302:
8114:
7780:
7566:
7452:
7388:
6800:
6531:
6502:
6464:
6243:
6142:
6056:
5927:
5898:
5856:
5770:
5586:
5557:
5503:
5421:
5347:
5280:
5253:
5221:
5199:
5132:
5054:
4999:
4912:
4879:
4849:would similarly make
4844:
4811:
4778:
4727:
4060:
4031:
4009:
3974:
3785:
3747:
3725:
3696:
3674:
3645:
3612:
3583:
3554:
3528:
3469:
3440:
3418:
3389:
3335:
3306:
3268:
2980:
2939:
2917:. The expression for
2912:
2880:
2671:
2590:
2427:
2385:
2356:
2336:
2304:
2195:, this means we take
2144:
2071:
2007:
1803:
1774:
1737:
1710:
1695:: find a sequence of
1677:
1562:
1421:
1351:
1270:
1234:
1198:
1142:
1076:
1046:
1013:
984:
951:
931:
905:
846:
824:
801:
736:
714:
692:
670:
645:
474:
449:
399:
339:
310:
280:
235:
213:
187:
162:
66:positive-semidefinite
22:
15440:. Dover Publishing.
14808:10.6028/jres.049.044
14592:
14563:
14543:
14292:
14254:
14183:
14015:
13988:
13931:
13839:
13819:is used, see above.
13756:
13663:
13541:
13394:
13356:
13334:
13312:
13283:
13252:
13221:
13199:
13153:
13092:
13068:
13062:forward substitution
13042:
13020:
12998:
12976:
12939:
12893:
12869:
12792:
12708:
12659:
12625:
12579:
12555:
12532:
12500:
12478:
12414:
12347:
12315:
12228:
12120:
12062:
11980:
11867:
11785:
11684:
11608:
11504:
11471:
11426:
11369:
11313:
11288:
11251:
11219:
11176:
11141:
11072:
10925:
10903:
10872:
10843:
10820:
10709:
10671:
10636:
10169:
10107:
10078:
10034:
10011:
9980:
9879:
9837:
9791:
9731:
9692:
9659:
9627:
9607:
9578:
9545:
9328:
9040:
8835:
8578:
8331:
8136:
7830:
7613:
7475:
7403:
7259:
6543:
6511:
6476:
6264:
6156:
6071:
5946:
5907:
5872:
5868:is chosen such that
5782:
5598:
5566:
5531:
5527:is chosen such that
5434:
5363:
5289:
5262:
5230:
5208:
5141:
5063:
5008:
4933:
4892:
4853:
4820:
4791:
4758:
4076:
4040:
4018:
3983:
3794:
3759:
3734:
3705:
3683:
3654:
3625:
3592:
3563:
3537:
3478:
3449:
3427:
3398:
3344:
3315:
3286:
2991:
2962:
2921:
2892:
2686:
2605:
2444:
2407:
2365:
2345:
2316:
2255:
2094:
2026:
1896:
1783:
1754:
1719:
1699:
1577:
1282:
1243:
1210:
1058:
1022:
993:
964:
940:
914:
855:
833:
813:
748:
723:
701:
679:
657:
489:
461:
411:
386:
319:
297:
267:
222:
196:
174:
136:
15603:Numerical stability
15457:Matrix Computations
15105:Matrix Computations
14780:Hestenes, Magnus R.
14228:conjugate transpose
14224:conjugate transpose
14218:(i.e., A' = A) and
13690:
13623:
13584:
13501:
13437:
11949:
11910:
11574:
11541:
10320:
10237:
9896:
9868:Convergence theorem
9110:
9077:
8648:
8615:
8122:We can now compute
7900:
7867:
6856:conjugate_gradient!
6782:
6704:
6564:
6440:
6353:
6291:
6030:
5992:
5746:
5713:
5668:
5635:
4783:is computed by the
4715: as the result
4568:
4529:
4297:
4264:
4165: as the result
3499:
3365:
3234:
3184:
2851:
2818:
2773:
2723:
2552:
2514:
2312:As observed above,
1400:
1330:
1303:
1264:
876:
93:energy minimization
70:iterative algorithm
14696:
14578:
14549:
14529:
14523:
14473:
14278:
14204:
14028:
14001:
13974:
13913:
13775:
13712:
13666:
13639:
13605:
13560:
13517:
13483:
13413:
13364:
13342:
13320:
13291:
13269:
13238:
13207:
13183:
13138:
13076:
13050:
13028:
13006:
12984:
12960:
12924:
12877:
12853:
12774:
12690:
12644:
12607:
12563:
12540:
12518:
12486:
12437:
12400:
12333:
12297:
12208:
12081:
12047:
11965:
11931:
11886:
11852:
11748:
11669:
11593:
11556:
11523:
11484:
11456:
11411:
11354:
11296:
11274:
11237:
11182:
11162:
11107:
11047:
10909:
10889:
10855:
10829:
10806:
10728:
10688:
10653:
10619:
10617:
10350:
10322:
10306:
10239:
10223:
10152:
10093:
10064:
10017:
10007:of maximal degree
9993:
9963:
9882:
9854:
9823:
9777:
9717:
9674:
9642:
9613:
9593:
9560:
9479:
9470:
9441:
9406:
9294:
9279:
9250:
9214:
9185:
9156:
9092:
9059:
8995:
8986:
8954:
8916:
8799:
8784:
8752:
8720:
8691:
8630:
8597:
8542:
8533:
8501:
8469:
8420:
8324:using the formula
8297:
8288:
8259:
8214:
8129:using the formula
8109:
8081:
8049:
8013:
7981:
7949:
7882:
7849:
7775:
7751:
7716:
7690:
7651:
7561:
7555:
7523:
7447:
7441:
7383:
7374:
7345:
7305:
7147:A_search_direction
7090:A_search_direction
7045:A_search_direction
6850:"""
6841:"""
6795:
6764:
6686:
6546:
6526:
6497:
6459:
6416:
6329:
6267:
6238:
6137:
6051:
6012:
5968:
5922:
5893:
5851:
5765:
5728:
5695:
5650:
5617:
5581:
5552:
5520:In the algorithm,
5498:
5416:
5342:
5275:
5248:
5216:
5194:
5127:
5049:
4994:
4907:
4874:
4839:
4806:
4787:method applied to
4773:
4722:
4720:
4550:
4505:
4458:
4279:
4246:
4132:
4055:
4026:
4004:
3969:
3780:
3742:
3720:
3691:
3669:
3640:
3607:
3578:
3549:
3523:
3481:
3464:
3435:
3413:
3384:
3347:
3330:
3301:
3263:
3261:
3216:
3166:
2975:
2934:
2907:
2875:
2833:
2800:
2755:
2705:
2666:
2585:
2534:
2496:
2492:
2422:
2380:
2351:
2331:
2299:
2174:. The gradient of
2139:
2066:
2002:
1926:
1888:quadratic function
1798:
1769:
1732:
1705:
1672:
1557:
1382:
1312:
1285:
1265:
1246:
1229:
1193:
1041:
1008:
979:
946:
926:
900:
858:
841:
819:
796:
731:
709:
687:
665:
640:
469:
444:
394:
334:
305:
275:
230:
208:
192:, where the known
182:
157:
58:numerical solution
42:
15729:
15728:
15520:978-1-61197-785-1
15495:978-0-89871-534-7
15466:978-1-4214-0794-4
15447:978-0-486-43227-4
15428:978-0-471-50023-0
15308:10.1137/060675290
15114:978-1-4214-0794-4
15054:978-3-319-28483-5
15029:978-0-89871-534-7
14971:10.1002/cpa.22081
14929:978-0-89871-396-1
14786:(December 1952).
14581:{\textstyle Wx=b}
14552:{\displaystyle W}
14458:
14443:
14404:
14389:
14369:
14354:
14334:
14319:
13965:
13943:
13637:
13515:
12269:
12182:
11963:
11807:
11591:
11376:
11185:{\displaystyle A}
11160:
11105:
11020:
11014:
11013:
10982:
10973:
10947:
10746:
10727:
10581:
10573:
10572:
10491:
10477:
10468:
10442:
10375:
10324:
10295:
10212:
10020:{\displaystyle k}
9953:
9932:
9926:
9907:
9616:{\displaystyle A}
9307:Finally, we find
9286:
9127:
8791:
8662:
8437:
8231:
8101:
8088:
7917:
7519:
7502:
7240:Numerical example
7123:# Update residual
7096:# Update solution
6762:
6684:
6414:
6327:
6250:the numerator of
6194:
6150:and equivalently
6049:
5763:
5687:
5562:is orthogonal to
5226:since the vector
4716:
4693:
4683:
4582:
4481:
4457:
4314:
4222:
4166:
4149:
4131:
3394:, for i ≠ j. And
3311:is orthogonal to
3253:
3140:
2870:
2792:
2571:
2477:
2354:{\displaystyle f}
1925:
1708:{\displaystyle n}
1667:
949:{\displaystyle P}
822:{\displaystyle n}
254:positive-definite
91:problems such as
15754:
15747:Gradient methods
15639:Matrix splitting
15577:
15570:
15563:
15554:
15553:
15549:
15524:
15499:
15487:
15479:
15470:
15451:
15432:
15420:
15400:
15399:
15393:
15385:
15383:
15382:
15367:
15361:
15360:
15358:
15349:
15343:
15329:
15320:
15319:
15301:
15281:
15275:
15274:
15264:
15254:
15230:
15224:
15223:
15213:
15204:(4): 1444–1460.
15193:
15187:
15186:
15176:
15156:
15150:
15149:
15125:
15119:
15118:
15100:
15094:
15093:
15091:
15090:
15084:
15073:
15067:
15066:
15040:
15034:
15033:
15009:
15003:
15002:
15000:
14989:
14983:
14982:
14964:
14955:(5): 1085–1136.
14940:
14934:
14933:
14904:
14893:
14892:
14879:
14866:
14865:
14844:
14838:
14837:
14835:
14833:2060/19710026200
14819:
14813:
14812:
14810:
14792:
14776:
14705:
14703:
14702:
14697:
14689:
14688:
14679:
14678:
14653:
14652:
14640:
14626:
14625:
14616:
14615:
14587:
14585:
14584:
14579:
14558:
14556:
14555:
14550:
14538:
14536:
14535:
14530:
14528:
14527:
14478:
14477:
14459:
14454:
14451:
14450:
14449:
14448:
14444:
14439:
14436:
14435:
14434:
14433:
14432:
14429:
14428:
14422:
14421:
14420:
14417:
14405:
14400:
14397:
14396:
14393:
14392:
14390:
14385:
14370:
14365:
14362:
14359:
14358:
14357:
14355:
14350:
14335:
14330:
14325:
14324:
14323:
14322:
14320:
14315:
14287:
14285:
14284:
14279:
14213:
14211:
14210:
14205:
14203:
14195:
14190:
14169:condition number
14144:) is equal to κ(
14135:condition number
14086:. The result is
14056:normal equations
14037:
14035:
14034:
14029:
14027:
14026:
14010:
14008:
14007:
14002:
14000:
13999:
13983:
13981:
13980:
13975:
13967:
13966:
13958:
13945:
13944:
13936:
13922:
13920:
13919:
13914:
13909:
13908:
13881:
13880:
13813:steepest descent
13810:
13804:
13798:
13791:steepest descent
13784:
13782:
13781:
13776:
13768:
13767:
13728:
13721:
13719:
13718:
13713:
13702:
13701:
13696:
13689:
13688:
13682:
13671:
13648:
13646:
13645:
13640:
13638:
13636:
13635:
13634:
13629:
13622:
13621:
13615:
13610:
13603:
13602:
13601:
13590:
13583:
13582:
13576:
13565:
13558:
13553:
13552:
13526:
13524:
13523:
13518:
13516:
13514:
13513:
13512:
13507:
13500:
13499:
13493:
13488:
13481:
13480:
13476:
13475:
13474:
13469:
13460:
13459:
13448:
13436:
13435:
13429:
13418:
13411:
13406:
13405:
13373:
13371:
13370:
13365:
13363:
13351:
13349:
13348:
13343:
13341:
13329:
13327:
13326:
13321:
13319:
13300:
13298:
13297:
13292:
13290:
13278:
13276:
13275:
13270:
13268:
13267:
13266:
13260:
13247:
13245:
13244:
13239:
13237:
13236:
13235:
13229:
13216:
13214:
13213:
13208:
13206:
13192:
13190:
13189:
13184:
13182:
13177:
13176:
13175:
13169:
13160:
13147:
13145:
13144:
13139:
13137:
13132:
13131:
13130:
13120:
13119:
13111:
13099:
13085:
13083:
13082:
13077:
13075:
13059:
13057:
13056:
13051:
13049:
13037:
13035:
13034:
13029:
13027:
13015:
13013:
13012:
13007:
13005:
12993:
12991:
12990:
12985:
12983:
12969:
12967:
12966:
12961:
12959:
12954:
12946:
12933:
12931:
12930:
12925:
12923:
12918:
12917:
12909:
12900:
12886:
12884:
12883:
12878:
12876:
12862:
12860:
12859:
12854:
12852:
12847:
12846:
12838:
12832:
12831:
12830:
12820:
12819:
12811:
12799:
12783:
12781:
12780:
12775:
12773:
12772:
12764:
12758:
12757:
12756:
12746:
12745:
12737:
12725:
12724:
12716:
12699:
12697:
12696:
12691:
12689:
12684:
12683:
12675:
12666:
12653:
12651:
12650:
12645:
12643:
12635:
12616:
12614:
12613:
12608:
12606:
12605:
12604:
12598:
12586:
12572:
12570:
12569:
12564:
12562:
12549:
12547:
12546:
12541:
12539:
12527:
12525:
12524:
12519:
12517:
12516:
12508:
12495:
12493:
12492:
12487:
12485:
12446:
12444:
12443:
12438:
12436:
12431:
12430:
12422:
12409:
12407:
12406:
12401:
12399:
12398:
12397:
12387:
12386:
12378:
12369:
12364:
12363:
12355:
12342:
12340:
12339:
12334:
12332:
12331:
12323:
12306:
12304:
12303:
12298:
12293:
12288:
12287:
12286:
12280:
12271:
12270:
12262:
12255:
12247:
12246:
12245:
12239:
12217:
12215:
12214:
12209:
12207:
12202:
12201:
12193:
12184:
12183:
12175:
12172:
12171:
12170:
12160:
12159:
12151:
12142:
12137:
12136:
12128:
12090:
12088:
12087:
12082:
12056:
12054:
12053:
12048:
12046:
12045:
12040:
12034:
12033:
12021:
12020:
12009:
12000:
11999:
11988:
11974:
11972:
11971:
11966:
11964:
11962:
11961:
11960:
11955:
11948:
11947:
11941:
11936:
11929:
11928:
11927:
11916:
11909:
11908:
11902:
11891:
11884:
11879:
11878:
11861:
11859:
11858:
11853:
11851:
11850:
11839:
11830:
11829:
11818:
11812:
11805:
11804:
11757:
11755:
11754:
11749:
11747:
11746:
11741:
11732:
11731:
11719:
11718:
11713:
11704:
11703:
11692:
11678:
11676:
11675:
11670:
11668:
11667:
11662:
11656:
11655:
11643:
11642:
11637:
11628:
11627:
11616:
11602:
11600:
11599:
11594:
11592:
11590:
11589:
11588:
11583:
11573:
11572:
11566:
11561:
11554:
11553:
11552:
11547:
11540:
11539:
11533:
11528:
11521:
11516:
11515:
11493:
11491:
11490:
11485:
11465:
11463:
11462:
11457:
11455:
11454:
11449:
11440:
11439:
11434:
11420:
11418:
11417:
11412:
11410:
11409:
11404:
11395:
11394:
11389:
11383:
11378:
11377:
11374:
11363:
11361:
11360:
11355:
11353:
11352:
11347:
11335:
11327:
11326:
11321:
11305:
11303:
11302:
11297:
11295:
11283:
11281:
11280:
11275:
11273:
11268:
11267:
11259:
11246:
11244:
11243:
11238:
11236:
11235:
11227:
11191:
11189:
11188:
11183:
11171:
11169:
11168:
11163:
11161:
11156:
11145:
11116:
11114:
11113:
11108:
11106:
11104:
11100:
11085:
11056:
11054:
11053:
11048:
11033:
11021:
11018:
11015:
11009:
10998:
10994:
10983:
10981:
10974:
10969:
10958:
10955:
10948:
10943:
10932:
10929:
10918:
10916:
10915:
10910:
10898:
10896:
10895:
10890:
10885:
10864:
10862:
10861:
10856:
10838:
10836:
10835:
10830:
10815:
10813:
10812:
10807:
10805:
10801:
10800:
10799:
10787:
10786:
10785:
10779:
10775:
10774:
10769:
10747:
10742:
10731:
10729:
10720:
10700:condition number
10697:
10695:
10694:
10689:
10684:
10662:
10660:
10659:
10654:
10649:
10628:
10626:
10625:
10620:
10618:
10610:
10609:
10608:
10602:
10598:
10597:
10592:
10579:
10578:
10574:
10568:
10557:
10556:
10545:
10524:
10520:
10519:
10518:
10512:
10508:
10507:
10502:
10489:
10488:
10487:
10482:
10478:
10476:
10469:
10464:
10453:
10450:
10443:
10438:
10427:
10424:
10408:
10404:
10403:
10402:
10396:
10392:
10391:
10386:
10373:
10372:
10355:
10349:
10345:
10321:
10319:
10314:
10288:
10284:
10283:
10282:
10276:
10272:
10271:
10270:
10265:
10256:
10238:
10236:
10231:
10204:
10203:
10202:
10196:
10192:
10191:
10186:
10161:
10159:
10158:
10153:
10151:
10150:
10145:
10136:
10135:
10130:
10121:
10120:
10115:
10102:
10100:
10099:
10094:
10092:
10091:
10086:
10073:
10071:
10070:
10065:
10063:
10062:
10057:
10053:
10052:
10047:
10026:
10024:
10023:
10018:
10002:
10000:
9999:
9994:
9992:
9991:
9972:
9970:
9969:
9964:
9958:
9954:
9951:
9930:
9924:
9923:
9922:
9905:
9895:
9890:
9863:
9861:
9860:
9855:
9850:
9833:is smaller than
9832:
9830:
9829:
9824:
9819:
9814:
9813:
9805:
9786:
9784:
9783:
9778:
9767:
9759:
9748:
9747:
9739:
9726:
9724:
9723:
9718:
9710:
9702:
9683:
9681:
9680:
9675:
9651:
9649:
9648:
9643:
9622:
9620:
9619:
9614:
9602:
9600:
9599:
9594:
9573:condition number
9569:
9567:
9566:
9561:
9559:
9558:
9553:
9539:iterative method
9517:= 2 iterations (
9488:
9486:
9485:
9480:
9475:
9474:
9446:
9445:
9411:
9410:
9382:
9381:
9376:
9370:
9369:
9357:
9356:
9351:
9342:
9341:
9336:
9303:
9301:
9300:
9295:
9287:
9285:
9284:
9283:
9255:
9254:
9219:
9218:
9191:
9190:
9189:
9161:
9160:
9133:
9128:
9126:
9125:
9124:
9119:
9109:
9108:
9102:
9097:
9090:
9089:
9088:
9083:
9076:
9075:
9069:
9064:
9057:
9052:
9051:
9032:
9016:
9004:
9002:
9001:
8996:
8991:
8990:
8959:
8958:
8921:
8920:
8889:
8888:
8883:
8877:
8876:
8864:
8863:
8858:
8849:
8848:
8843:
8820:
8808:
8806:
8805:
8800:
8792:
8790:
8789:
8788:
8757:
8756:
8726:
8725:
8724:
8696:
8695:
8668:
8663:
8661:
8660:
8659:
8654:
8647:
8646:
8640:
8635:
8628:
8627:
8626:
8621:
8614:
8613:
8607:
8602:
8595:
8590:
8589:
8563:
8551:
8549:
8548:
8543:
8538:
8537:
8506:
8505:
8474:
8473:
8438:
8430:
8425:
8424:
8390:
8389:
8384:
8378:
8373:
8372:
8360:
8359:
8354:
8345:
8344:
8339:
8306:
8304:
8303:
8298:
8293:
8292:
8264:
8263:
8232:
8224:
8219:
8218:
8190:
8189:
8184:
8178:
8177:
8165:
8164:
8159:
8150:
8149:
8144:
8118:
8116:
8115:
8110:
8102:
8094:
8089:
8087:
8086:
8085:
8054:
8053:
8018:
8017:
7987:
7986:
7985:
7954:
7953:
7923:
7918:
7916:
7915:
7914:
7909:
7899:
7898:
7892:
7887:
7880:
7879:
7878:
7873:
7866:
7865:
7859:
7854:
7847:
7842:
7841:
7822:
7784:
7782:
7781:
7776:
7771:
7770:
7765:
7756:
7755:
7721:
7720:
7695:
7694:
7656:
7655:
7627:
7626:
7621:
7581:associated with
7570:
7568:
7567:
7562:
7560:
7559:
7546:
7528:
7527:
7520:
7512:
7507:
7503:
7495:
7482:
7456:
7454:
7453:
7448:
7446:
7445:
7417:
7416:
7411:
7392:
7390:
7389:
7384:
7379:
7378:
7350:
7349:
7342:
7341:
7328:
7327:
7310:
7309:
7271:
7266:
7235:
7232:
7229:
7226:
7223:
7220:
7217:
7214:
7211:
7210:search_direction
7208:
7205:
7202:
7199:
7196:
7193:
7190:
7187:
7184:
7181:
7178:
7175:
7174:search_direction
7172:
7169:
7166:
7163:
7160:
7157:
7154:
7151:
7148:
7145:
7142:
7139:
7136:
7133:
7130:
7127:
7124:
7121:
7120:search_direction
7118:
7115:
7112:
7109:
7106:
7103:
7100:
7097:
7094:
7091:
7088:
7085:
7082:
7081:search_direction
7079:
7076:
7073:
7070:
7067:
7064:
7061:
7058:
7057:search_direction
7055:
7052:
7049:
7046:
7043:
7040:
7037:
7034:
7031:
7028:
7025:
7022:
7019:
7016:
7013:
7010:
7007:
7004:
7001:
6998:
6995:
6992:
6989:
6986:
6983:
6980:
6977:
6974:
6971:
6968:
6965:
6962:
6959:
6956:
6953:
6952:search_direction
6950:
6947:
6944:
6941:
6938:
6935:
6932:
6929:
6926:
6923:
6920:
6917:
6914:
6911:
6908:
6905:
6902:
6899:
6896:
6893:
6890:
6887:
6884:
6881:
6878:
6875:
6872:
6869:
6866:
6863:
6860:
6857:
6854:
6851:
6848:
6845:
6842:
6832:Example code in
6827:
6820:
6804:
6802:
6801:
6796:
6794:
6793:
6788:
6781:
6780:
6774:
6769:
6763:
6761:
6760:
6748:
6740:
6739:
6728:
6719:
6718:
6713:
6703:
6702:
6696:
6691:
6685:
6683:
6682:
6670:
6665:
6664:
6659:
6653:
6648:
6647:
6646:
6636:
6635:
6624:
6618:
6617:
6599:
6598:
6593:
6581:
6580:
6575:
6569:
6563:
6562:
6556:
6551:
6535:
6533:
6532:
6527:
6525:
6524:
6519:
6506:
6504:
6503:
6498:
6496:
6495:
6484:
6468:
6466:
6465:
6460:
6458:
6457:
6446:
6439:
6438:
6432:
6421:
6415:
6413:
6412:
6400:
6389:
6388:
6377:
6368:
6367:
6362:
6352:
6351:
6345:
6334:
6328:
6326:
6325:
6313:
6308:
6307:
6302:
6296:
6290:
6289:
6283:
6272:
6257:is rewritten as
6256:
6247:
6245:
6244:
6239:
6231:
6230:
6219:
6210:
6209:
6204:
6195:
6193:
6192:
6180:
6175:
6174:
6169:
6163:
6146:
6144:
6143:
6138:
6136:
6135:
6130:
6124:
6119:
6118:
6106:
6105:
6100:
6091:
6090:
6079:
6060:
6058:
6057:
6052:
6050:
6048:
6047:
6046:
6041:
6035:
6029:
6028:
6022:
6017:
6010:
6009:
6008:
6003:
5997:
5991:
5990:
5984:
5973:
5966:
5958:
5957:
5938:
5931:
5929:
5928:
5923:
5921:
5920:
5915:
5903:is conjugate to
5902:
5900:
5899:
5894:
5892:
5891:
5880:
5867:
5860:
5858:
5857:
5852:
5850:
5849:
5844:
5838:
5837:
5832:
5823:
5822:
5811:
5802:
5801:
5790:
5774:
5772:
5771:
5766:
5764:
5762:
5761:
5760:
5755:
5745:
5744:
5738:
5733:
5726:
5725:
5724:
5719:
5712:
5711:
5705:
5700:
5693:
5688:
5686:
5685:
5684:
5679:
5673:
5667:
5666:
5660:
5655:
5648:
5647:
5646:
5641:
5634:
5633:
5627:
5622:
5615:
5610:
5609:
5590:
5588:
5587:
5582:
5580:
5579:
5574:
5561:
5559:
5558:
5553:
5551:
5550:
5539:
5526:
5507:
5505:
5504:
5499:
5497:
5496:
5491:
5482:
5481:
5469:
5468:
5463:
5454:
5453:
5442:
5425:
5423:
5422:
5417:
5415:
5414:
5403:
5391:
5383:
5382:
5371:
5351:
5349:
5348:
5343:
5341:
5340:
5329:
5317:
5309:
5308:
5297:
5284:
5282:
5281:
5276:
5274:
5273:
5257:
5255:
5254:
5249:
5247:
5246:
5241:
5225:
5223:
5222:
5217:
5215:
5203:
5201:
5200:
5195:
5193:
5192:
5181:
5169:
5161:
5160:
5149:
5136:
5134:
5133:
5128:
5126:
5125:
5120:
5111:
5110:
5098:
5097:
5092:
5083:
5082:
5071:
5058:
5056:
5055:
5050:
5048:
5047:
5042:
5030:
5022:
5021:
5016:
5003:
5001:
5000:
4995:
4993:
4992:
4987:
4981:
4980:
4968:
4967:
4962:
4953:
4952:
4941:
4916:
4914:
4913:
4908:
4906:
4905:
4900:
4886:gradient descent
4884:computed by the
4883:
4881:
4880:
4875:
4873:
4872:
4861:
4848:
4846:
4845:
4840:
4832:
4831:
4815:
4813:
4812:
4807:
4805:
4804:
4799:
4785:gradient descent
4782:
4780:
4779:
4774:
4772:
4771:
4766:
4741:
4731:
4729:
4728:
4723:
4721:
4717:
4714:
4712:
4711:
4700:
4694:
4691:
4688:
4684:
4681:
4678:
4657:
4653:
4652:
4647:
4641:
4640:
4628:
4627:
4616:
4607:
4606:
4595:
4587:
4583:
4581:
4580:
4579:
4574:
4567:
4566:
4560:
4555:
4548:
4547:
4546:
4535:
4528:
4527:
4521:
4510:
4503:
4498:
4497:
4486:
4482:
4479:
4477:
4476:
4465:
4459:
4455:
4450:
4446:
4445:
4440:
4431:
4430:
4418:
4417:
4412:
4403:
4402:
4391:
4383:
4379:
4378:
4373:
4367:
4366:
4354:
4353:
4348:
4339:
4338:
4327:
4319:
4315:
4313:
4312:
4311:
4306:
4296:
4295:
4289:
4284:
4277:
4276:
4275:
4270:
4263:
4262:
4256:
4251:
4244:
4239:
4238:
4227:
4223:
4220:
4217:
4203:
4199:
4198:
4193:
4184:
4183:
4178:
4171:
4167:
4164:
4162:
4161:
4156:
4150:
4147:
4145:
4144:
4139:
4133:
4129:
4125:
4121:
4120:
4115:
4103:
4095:
4094:
4089:
4082:
4064:
4062:
4061:
4056:
4054:
4053:
4048:
4035:
4033:
4032:
4027:
4025:
4013:
4011:
4010:
4005:
4003:
3995:
3990:
3978:
3976:
3975:
3970:
3968:
3967:
3963:
3962:
3958:
3954:
3953:
3879:
3878:
3851:
3850:
3849:
3848:
3843:
3830:
3806:
3805:
3789:
3787:
3786:
3781:
3773:
3772:
3767:
3751:
3749:
3748:
3743:
3741:
3729:
3727:
3726:
3721:
3719:
3718:
3713:
3700:
3698:
3697:
3692:
3690:
3678:
3676:
3675:
3670:
3668:
3667:
3662:
3649:
3647:
3646:
3641:
3639:
3638:
3633:
3616:
3614:
3613:
3608:
3606:
3605:
3600:
3587:
3585:
3584:
3579:
3577:
3576:
3571:
3558:
3556:
3555:
3550:
3532:
3530:
3529:
3524:
3516:
3515:
3510:
3504:
3498:
3497:
3491:
3486:
3473:
3471:
3470:
3465:
3463:
3462:
3457:
3444:
3442:
3441:
3436:
3434:
3422:
3420:
3419:
3414:
3412:
3411:
3406:
3393:
3391:
3390:
3385:
3377:
3376:
3371:
3364:
3363:
3357:
3352:
3339:
3337:
3336:
3331:
3329:
3328:
3323:
3310:
3308:
3307:
3302:
3300:
3299:
3294:
3272:
3270:
3269:
3264:
3262:
3254:
3252:
3251:
3250:
3245:
3239:
3233:
3232:
3226:
3221:
3214:
3210:
3209:
3204:
3192:
3183:
3182:
3176:
3171:
3164:
3159:
3158:
3141:
3133:
3124:
3123:
3111:
3096:
3095:
3074:
3073:
3068:
3062:
3061:
3049:
3048:
3043:
3021:
3020:
3009:
2984:
2982:
2981:
2976:
2974:
2973:
2943:
2941:
2940:
2935:
2933:
2932:
2916:
2914:
2913:
2908:
2906:
2905:
2900:
2884:
2882:
2881:
2876:
2871:
2869:
2868:
2867:
2862:
2856:
2850:
2849:
2843:
2838:
2831:
2830:
2829:
2824:
2817:
2816:
2810:
2805:
2798:
2793:
2791:
2790:
2789:
2784:
2778:
2772:
2771:
2765:
2760:
2753:
2749:
2748:
2743:
2731:
2722:
2721:
2715:
2710:
2703:
2698:
2697:
2675:
2673:
2672:
2667:
2665:
2664:
2659:
2653:
2652:
2640:
2639:
2634:
2625:
2624:
2613:
2594:
2592:
2591:
2586:
2584:
2583:
2578:
2572:
2570:
2569:
2568:
2563:
2557:
2551:
2550:
2544:
2539:
2532:
2531:
2530:
2525:
2519:
2513:
2512:
2506:
2501:
2494:
2491:
2473:
2472:
2467:
2458:
2457:
2452:
2431:
2429:
2428:
2423:
2421:
2420:
2415:
2392:gradient descent
2389:
2387:
2386:
2381:
2379:
2378:
2373:
2360:
2358:
2357:
2352:
2340:
2338:
2337:
2332:
2330:
2329:
2324:
2308:
2306:
2305:
2300:
2295:
2294:
2289:
2277:
2269:
2268:
2263:
2187:
2148:
2146:
2145:
2140:
2134:
2126:
2121:
2110:
2075:
2073:
2072:
2067:
2061:
2047:
2033:
2011:
2009:
2008:
2003:
1997:
1996:
1991:
1982:
1973:
1968:
1967:
1966:
1960:
1951:
1946:
1941:
1940:
1939:
1933:
1927:
1918:
1909:
1885:
1876:
1845:
1832:
1823:
1807:
1805:
1804:
1799:
1797:
1796:
1791:
1778:
1776:
1775:
1770:
1768:
1767:
1762:
1741:
1739:
1738:
1733:
1731:
1730:
1714:
1712:
1711:
1706:
1694:
1681:
1679:
1678:
1673:
1668:
1666:
1665:
1664:
1663:
1653:
1652:
1647:
1638:
1637:
1632:
1622:
1618:
1610:
1609:
1604:
1594:
1589:
1588:
1566:
1564:
1563:
1558:
1556:
1555:
1554:
1548:
1544:
1543:
1542:
1537:
1528:
1527:
1522:
1510:
1509:
1497:
1496:
1495:
1489:
1485:
1484:
1483:
1478:
1469:
1468:
1463:
1451:
1450:
1440:
1435:
1417:
1416:
1411:
1405:
1399:
1398:
1392:
1387:
1381:
1380:
1370:
1365:
1347:
1346:
1341:
1335:
1329:
1328:
1322:
1317:
1308:
1302:
1301:
1295:
1290:
1274:
1272:
1271:
1266:
1263:
1262:
1256:
1251:
1239:with the vector
1238:
1236:
1235:
1230:
1228:
1220:
1202:
1200:
1199:
1194:
1189:
1188:
1183:
1177:
1172:
1171:
1161:
1156:
1138:
1137:
1132:
1126:
1118:
1117:
1112:
1106:
1105:
1095:
1090:
1072:
1071:
1066:
1050:
1048:
1047:
1042:
1040:
1032:
1017:
1015:
1014:
1009:
1007:
1006:
1001:
988:
986:
985:
980:
978:
977:
972:
955:
953:
952:
947:
935:
933:
932:
927:
909:
907:
906:
901:
893:
892:
887:
881:
875:
874:
868:
863:
850:
848:
847:
842:
840:
828:
826:
825:
820:
805:
803:
802:
797:
792:
791:
786:
771:
770:
765:
740:
738:
737:
732:
730:
719:is conjugate to
718:
716:
715:
710:
708:
696:
694:
693:
688:
686:
675:is conjugate to
674:
672:
671:
666:
664:
649:
647:
646:
641:
633:
628:
620:
606:
601:
600:
599:
593:
584:
570:
562:
557:
546:
545:
544:
534:
526:
515:
510:
505:
504:
503:
497:
478:
476:
475:
470:
468:
453:
451:
450:
445:
437:
432:
427:
426:
425:
419:
403:
401:
400:
395:
393:
343:
341:
340:
335:
333:
332:
327:
314:
312:
311:
306:
304:
284:
282:
281:
276:
274:
239:
237:
236:
231:
229:
217:
215:
214:
209:
191:
189:
188:
183:
181:
166:
164:
163:
158:
156:
148:
143:
72:, applicable to
40: = 2).
25:gradient descent
15762:
15761:
15757:
15756:
15755:
15753:
15752:
15751:
15732:
15731:
15730:
15725:
15684:
15680:Multiprocessing
15648:
15644:Sparse problems
15607:
15586:
15581:
15534:
15531:
15521:
15496:
15467:
15448:
15429:
15409:
15407:Further reading
15404:
15403:
15387:
15386:
15380:
15378:
15368:
15364:
15356:
15350:
15346:
15330:
15323:
15282:
15278:
15231:
15227:
15194:
15190:
15157:
15153:
15146:10.1137/0906018
15126:
15122:
15115:
15101:
15097:
15088:
15086:
15082:
15074:
15070:
15055:
15041:
15037:
15030:
15010:
15006:
14998:
14990:
14986:
14941:
14937:
14930:
14908:Greenbaum, Anne
14905:
14896:
14880:
14869:
14862:
14848:Speiser, Ambros
14845:
14841:
14820:
14816:
14790:
14784:Stiefel, Eduard
14777:
14773:
14768:
14763:
14754:Preconditioning
14744:Krylov subspace
14712:
14684:
14680:
14674:
14670:
14648:
14644:
14636:
14621:
14617:
14611:
14607:
14593:
14590:
14589:
14564:
14561:
14560:
14544:
14541:
14540:
14522:
14521:
14515:
14514:
14508:
14507:
14501:
14500:
14490:
14489:
14472:
14471:
14460:
14453:
14446:
14445:
14438:
14430:
14427:
14418:
14416:
14411:
14406:
14399:
14394:
14391:
14384:
14382:
14371:
14364:
14360:
14356:
14349:
14347:
14336:
14329:
14326:
14321:
14314:
14312:
14302:
14301:
14293:
14290:
14289:
14255:
14252:
14251:
14248:
14240:
14199:
14191:
14186:
14184:
14181:
14180:
14177:
14082:matrix for any
14078:is a symmetric
14044:
14022:
14018:
14016:
14013:
14012:
13995:
13991:
13989:
13986:
13985:
13984:The quantities
13957:
13956:
13935:
13934:
13932:
13929:
13928:
13904:
13900:
13876:
13872:
13840:
13837:
13836:
13825:
13806:
13800:
13794:
13763:
13759:
13757:
13754:
13753:
13750:
13742:locally optimal
13731:round-off error
13727:
13723:
13697:
13692:
13691:
13684:
13683:
13672:
13667:
13664:
13661:
13660:
13630:
13625:
13624:
13617:
13616:
13611:
13606:
13604:
13591:
13586:
13585:
13578:
13577:
13566:
13561:
13559:
13557:
13548:
13544:
13542:
13539:
13538:
13532:Fletcher–Reeves
13530:instead of the
13508:
13503:
13502:
13495:
13494:
13489:
13484:
13482:
13470:
13465:
13464:
13449:
13444:
13443:
13442:
13438:
13431:
13430:
13419:
13414:
13412:
13410:
13401:
13397:
13395:
13392:
13391:
13380:
13359:
13357:
13354:
13353:
13337:
13335:
13332:
13331:
13315:
13313:
13310:
13309:
13286:
13284:
13281:
13280:
13262:
13261:
13256:
13255:
13253:
13250:
13249:
13248:are known, and
13231:
13230:
13225:
13224:
13222:
13219:
13218:
13202:
13200:
13197:
13196:
13178:
13171:
13170:
13165:
13164:
13156:
13154:
13151:
13150:
13133:
13126:
13125:
13121:
13112:
13107:
13106:
13095:
13093:
13090:
13089:
13071:
13069:
13066:
13065:
13045:
13043:
13040:
13039:
13023:
13021:
13018:
13017:
13016:and known, and
13001:
12999:
12996:
12995:
12979:
12977:
12974:
12973:
12955:
12950:
12942:
12940:
12937:
12936:
12919:
12910:
12905:
12904:
12896:
12894:
12891:
12890:
12872:
12870:
12867:
12866:
12848:
12839:
12834:
12833:
12826:
12825:
12821:
12812:
12807:
12806:
12795:
12793:
12790:
12789:
12765:
12760:
12759:
12752:
12751:
12747:
12738:
12733:
12732:
12717:
12712:
12711:
12709:
12706:
12705:
12685:
12676:
12671:
12670:
12662:
12660:
12657:
12656:
12639:
12628:
12626:
12623:
12622:
12600:
12599:
12591:
12590:
12582:
12580:
12577:
12576:
12558:
12556:
12553:
12552:
12535:
12533:
12530:
12529:
12509:
12504:
12503:
12501:
12498:
12497:
12481:
12479:
12476:
12475:
12472:
12432:
12423:
12418:
12417:
12415:
12412:
12411:
12393:
12392:
12388:
12379:
12374:
12373:
12365:
12356:
12351:
12350:
12348:
12345:
12344:
12324:
12319:
12318:
12316:
12313:
12312:
12289:
12282:
12281:
12276:
12275:
12261:
12260:
12251:
12241:
12240:
12232:
12231:
12229:
12226:
12225:
12203:
12194:
12189:
12188:
12174:
12173:
12166:
12165:
12161:
12152:
12147:
12146:
12138:
12129:
12124:
12123:
12121:
12118:
12117:
12109:
12063:
12060:
12059:
12041:
12036:
12035:
12029:
12025:
12010:
12005:
12004:
11989:
11984:
11983:
11981:
11978:
11977:
11956:
11951:
11950:
11943:
11942:
11937:
11932:
11930:
11917:
11912:
11911:
11904:
11903:
11892:
11887:
11885:
11883:
11874:
11870:
11868:
11865:
11864:
11840:
11835:
11834:
11819:
11814:
11813:
11808:
11788:
11786:
11783:
11782:
11771:
11742:
11734:
11733:
11727:
11723:
11714:
11709:
11708:
11693:
11688:
11687:
11685:
11682:
11681:
11663:
11658:
11657:
11651:
11647:
11638:
11633:
11632:
11617:
11612:
11611:
11609:
11606:
11605:
11584:
11576:
11575:
11568:
11567:
11562:
11557:
11555:
11548:
11543:
11542:
11535:
11534:
11529:
11524:
11522:
11520:
11511:
11507:
11505:
11502:
11501:
11472:
11469:
11468:
11450:
11445:
11444:
11435:
11430:
11429:
11427:
11424:
11423:
11405:
11400:
11399:
11390:
11385:
11384:
11379:
11373:
11372:
11370:
11367:
11366:
11348:
11340:
11339:
11331:
11322:
11317:
11316:
11314:
11311:
11310:
11291:
11289:
11286:
11285:
11269:
11260:
11255:
11254:
11252:
11249:
11248:
11228:
11223:
11222:
11220:
11217:
11216:
11213:preconditioning
11211:In most cases,
11209:
11203:
11177:
11174:
11173:
11152:
11144:
11142:
11139:
11138:
11134:
11122:round-off error
11096:
11089:
11084:
11073:
11070:
11069:
11068:which scale as
11029:
11017:
11005:
10993:
10965:
10957:
10956:
10939:
10931:
10930:
10928:
10926:
10923:
10922:
10904:
10901:
10900:
10881:
10873:
10870:
10869:
10844:
10841:
10840:
10821:
10818:
10817:
10792:
10788:
10781:
10780:
10770:
10765:
10764:
10760:
10759:
10758:
10754:
10738:
10730:
10718:
10710:
10707:
10706:
10680:
10672:
10669:
10668:
10645:
10637:
10634:
10633:
10616:
10615:
10604:
10603:
10593:
10588:
10587:
10583:
10582:
10564:
10546:
10544:
10540:
10522:
10521:
10514:
10513:
10503:
10498:
10497:
10493:
10492:
10483:
10460:
10452:
10451:
10434:
10426:
10425:
10423:
10419:
10418:
10406:
10405:
10398:
10397:
10387:
10382:
10381:
10377:
10376:
10368:
10351:
10341:
10328:
10315:
10310:
10299:
10286:
10285:
10278:
10277:
10266:
10261:
10260:
10252:
10245:
10241:
10240:
10232:
10227:
10216:
10205:
10198:
10197:
10187:
10182:
10181:
10177:
10176:
10172:
10170:
10167:
10166:
10146:
10141:
10140:
10131:
10126:
10125:
10116:
10111:
10110:
10108:
10105:
10104:
10087:
10082:
10081:
10079:
10076:
10075:
10058:
10048:
10043:
10042:
10038:
10037:
10035:
10032:
10031:
10012:
10009:
10008:
9987:
9983:
9981:
9978:
9977:
9918:
9914:
9904:
9900:
9891:
9886:
9880:
9877:
9876:
9870:
9846:
9838:
9835:
9834:
9815:
9806:
9801:
9800:
9792:
9789:
9788:
9763:
9752:
9740:
9735:
9734:
9732:
9729:
9728:
9706:
9695:
9693:
9690:
9689:
9686:preconditioning
9660:
9657:
9656:
9628:
9625:
9624:
9608:
9605:
9604:
9579:
9576:
9575:
9554:
9549:
9548:
9546:
9543:
9542:
9531:round-off error
9527:
9512:
9505:
9498:
9469:
9468:
9462:
9461:
9451:
9450:
9440:
9439:
9433:
9432:
9419:
9418:
9405:
9404:
9398:
9397:
9387:
9386:
9377:
9372:
9371:
9365:
9361:
9352:
9347:
9346:
9337:
9332:
9331:
9329:
9326:
9325:
9320:
9313:
9278:
9277:
9271:
9270:
9257:
9256:
9249:
9248:
9243:
9237:
9236:
9231:
9221:
9220:
9213:
9212:
9207:
9194:
9193:
9192:
9184:
9183:
9177:
9176:
9163:
9162:
9155:
9154:
9149:
9136:
9135:
9134:
9132:
9120:
9112:
9111:
9104:
9103:
9098:
9093:
9091:
9084:
9079:
9078:
9071:
9070:
9065:
9060:
9058:
9056:
9047:
9043:
9041:
9038:
9037:
9031:
9025:
9023:
9015:
9009:
8985:
8984:
8978:
8977:
8964:
8963:
8953:
8952:
8943:
8942:
8929:
8928:
8915:
8914:
8908:
8907:
8894:
8893:
8884:
8879:
8878:
8872:
8868:
8859:
8854:
8853:
8844:
8839:
8838:
8836:
8833:
8832:
8827:
8819:
8813:
8783:
8782:
8773:
8772:
8759:
8758:
8751:
8750:
8742:
8729:
8728:
8727:
8719:
8718:
8712:
8711:
8698:
8697:
8690:
8689:
8684:
8671:
8670:
8669:
8667:
8655:
8650:
8649:
8642:
8641:
8636:
8631:
8629:
8622:
8617:
8616:
8609:
8608:
8603:
8598:
8596:
8594:
8585:
8581:
8579:
8576:
8575:
8570:
8562:
8556:
8532:
8531:
8525:
8524:
8511:
8510:
8500:
8499:
8490:
8489:
8476:
8475:
8468:
8467:
8462:
8456:
8455:
8450:
8440:
8439:
8429:
8419:
8418:
8409:
8408:
8395:
8394:
8385:
8380:
8379:
8374:
8368:
8364:
8355:
8350:
8349:
8340:
8335:
8334:
8332:
8329:
8328:
8323:
8316:
8287:
8286:
8280:
8279:
8269:
8268:
8258:
8257:
8248:
8247:
8234:
8233:
8223:
8213:
8212:
8206:
8205:
8195:
8194:
8185:
8180:
8179:
8173:
8169:
8160:
8155:
8154:
8145:
8140:
8139:
8137:
8134:
8133:
8128:
8093:
8080:
8079:
8070:
8069:
8056:
8055:
8048:
8047:
8042:
8036:
8035:
8030:
8020:
8019:
8012:
8011:
8003:
7990:
7989:
7988:
7980:
7979:
7970:
7969:
7956:
7955:
7948:
7947:
7939:
7926:
7925:
7924:
7922:
7910:
7902:
7901:
7894:
7893:
7888:
7883:
7881:
7874:
7869:
7868:
7861:
7860:
7855:
7850:
7848:
7846:
7837:
7833:
7831:
7828:
7827:
7821:
7815:
7810:
7801:
7794:
7766:
7761:
7760:
7750:
7749:
7740:
7739:
7726:
7725:
7715:
7714:
7708:
7707:
7697:
7696:
7689:
7688:
7683:
7677:
7676:
7671:
7661:
7660:
7650:
7649:
7643:
7642:
7632:
7631:
7622:
7617:
7616:
7614:
7611:
7610:
7605:
7594:
7587:
7580:
7554:
7553:
7547:
7544:
7543:
7533:
7532:
7522:
7521:
7511:
7508:
7505:
7504:
7494:
7487:
7486:
7478:
7476:
7473:
7472:
7466:
7440:
7439:
7433:
7432:
7422:
7421:
7412:
7407:
7406:
7404:
7401:
7400:
7373:
7372:
7366:
7365:
7355:
7354:
7344:
7343:
7337:
7333:
7330:
7329:
7323:
7319:
7312:
7311:
7304:
7303:
7298:
7292:
7291:
7286:
7276:
7275:
7267:
7262:
7260:
7257:
7256:
7242:
7237:
7236:
7233:
7230:
7227:
7224:
7221:
7218:
7215:
7212:
7209:
7206:
7203:
7200:
7197:
7194:
7191:
7188:
7185:
7182:
7179:
7176:
7173:
7170:
7167:
7164:
7161:
7158:
7155:
7152:
7149:
7146:
7143:
7140:
7137:
7134:
7131:
7128:
7125:
7122:
7119:
7116:
7113:
7110:
7107:
7104:
7101:
7098:
7095:
7092:
7089:
7086:
7083:
7080:
7077:
7074:
7071:
7068:
7065:
7062:
7059:
7056:
7053:
7050:
7047:
7044:
7041:
7038:
7035:
7032:
7029:
7026:
7023:
7020:
7017:
7014:
7011:
7008:
7005:
7002:
6999:
6996:
6993:
6990:
6987:
6984:
6981:
6978:
6975:
6972:
6969:
6966:
6963:
6960:
6957:
6954:
6951:
6948:
6945:
6942:
6939:
6936:
6933:
6930:
6927:
6924:
6921:
6918:
6915:
6912:
6909:
6906:
6903:
6900:
6897:
6894:
6891:
6888:
6885:
6882:
6879:
6876:
6873:
6870:
6867:
6864:
6861:
6858:
6855:
6852:
6849:
6846:
6843:
6840:
6837:
6826:
6822:
6818:
6816:
6789:
6784:
6783:
6776:
6775:
6770:
6765:
6756:
6752:
6747:
6729:
6724:
6723:
6714:
6709:
6708:
6698:
6697:
6692:
6687:
6678:
6674:
6669:
6660:
6655:
6654:
6649:
6642:
6641:
6637:
6625:
6620:
6619:
6607:
6603:
6594:
6589:
6588:
6576:
6571:
6570:
6565:
6558:
6557:
6552:
6547:
6544:
6541:
6540:
6520:
6515:
6514:
6512:
6509:
6508:
6485:
6480:
6479:
6477:
6474:
6473:
6447:
6442:
6441:
6434:
6433:
6422:
6417:
6408:
6404:
6399:
6378:
6373:
6372:
6363:
6358:
6357:
6347:
6346:
6335:
6330:
6321:
6317:
6312:
6303:
6298:
6297:
6292:
6285:
6284:
6273:
6268:
6265:
6262:
6261:
6255:
6251:
6220:
6215:
6214:
6205:
6200:
6199:
6188:
6184:
6179:
6170:
6165:
6164:
6159:
6157:
6154:
6153:
6131:
6126:
6125:
6120:
6114:
6110:
6101:
6096:
6095:
6080:
6075:
6074:
6072:
6069:
6068:
6042:
6037:
6036:
6031:
6024:
6023:
6018:
6013:
6011:
6004:
5999:
5998:
5993:
5986:
5985:
5974:
5969:
5967:
5965:
5953:
5949:
5947:
5944:
5943:
5937:
5933:
5916:
5911:
5910:
5908:
5905:
5904:
5881:
5876:
5875:
5873:
5870:
5869:
5866:
5862:
5845:
5840:
5839:
5833:
5828:
5827:
5812:
5807:
5806:
5791:
5786:
5785:
5783:
5780:
5779:
5756:
5748:
5747:
5740:
5739:
5734:
5729:
5727:
5720:
5715:
5714:
5707:
5706:
5701:
5696:
5694:
5692:
5680:
5675:
5674:
5669:
5662:
5661:
5656:
5651:
5649:
5642:
5637:
5636:
5629:
5628:
5623:
5618:
5616:
5614:
5605:
5601:
5599:
5596:
5595:
5575:
5570:
5569:
5567:
5564:
5563:
5540:
5535:
5534:
5532:
5529:
5528:
5525:
5521:
5518:
5510:rounding errors
5492:
5484:
5483:
5477:
5473:
5464:
5459:
5458:
5443:
5438:
5437:
5435:
5432:
5431:
5428:rounding errors
5404:
5396:
5395:
5387:
5372:
5367:
5366:
5364:
5361:
5360:
5354:round-off error
5330:
5322:
5321:
5313:
5298:
5293:
5292:
5290:
5287:
5286:
5269:
5265:
5263:
5260:
5259:
5242:
5234:
5233:
5231:
5228:
5227:
5211:
5209:
5206:
5205:
5182:
5174:
5173:
5165:
5150:
5145:
5144:
5142:
5139:
5138:
5121:
5113:
5112:
5106:
5102:
5093:
5088:
5087:
5072:
5067:
5066:
5064:
5061:
5060:
5043:
5035:
5034:
5026:
5017:
5012:
5011:
5009:
5006:
5005:
4988:
4983:
4982:
4976:
4972:
4963:
4958:
4957:
4942:
4937:
4936:
4934:
4931:
4930:
4927:
4919:round-off error
4901:
4896:
4895:
4893:
4890:
4889:
4862:
4857:
4856:
4854:
4851:
4850:
4827:
4823:
4821:
4818:
4817:
4800:
4795:
4794:
4792:
4789:
4788:
4767:
4762:
4761:
4759:
4756:
4755:
4752:
4740:
4736:
4719:
4718:
4713:
4701:
4696:
4695:
4690:
4686:
4685:
4680:
4676:
4675:
4655:
4654:
4648:
4643:
4642:
4636:
4632:
4617:
4612:
4611:
4596:
4591:
4590:
4585:
4584:
4575:
4570:
4569:
4562:
4561:
4556:
4551:
4549:
4536:
4531:
4530:
4523:
4522:
4511:
4506:
4504:
4502:
4493:
4489:
4484:
4483:
4478:
4466:
4461:
4460:
4453:
4448:
4447:
4441:
4433:
4432:
4426:
4422:
4413:
4408:
4407:
4392:
4387:
4386:
4381:
4380:
4374:
4369:
4368:
4362:
4358:
4349:
4344:
4343:
4328:
4323:
4322:
4317:
4316:
4307:
4299:
4298:
4291:
4290:
4285:
4280:
4278:
4271:
4266:
4265:
4258:
4257:
4252:
4247:
4245:
4243:
4234:
4230:
4225:
4224:
4219:
4215:
4214:
4201:
4200:
4194:
4189:
4188:
4179:
4174:
4173:
4169:
4168:
4163:
4157:
4152:
4151:
4146:
4140:
4135:
4134:
4127:
4123:
4122:
4116:
4108:
4107:
4099:
4090:
4085:
4084:
4079:
4077:
4074:
4073:
4049:
4044:
4043:
4041:
4038:
4037:
4021:
4019:
4016:
4015:
3999:
3991:
3986:
3984:
3981:
3980:
3943:
3939:
3917:
3913:
3874:
3870:
3857:
3853:
3852:
3844:
3839:
3838:
3831:
3811:
3810:
3801:
3797:
3795:
3792:
3791:
3768:
3763:
3762:
3760:
3757:
3756:
3737:
3735:
3732:
3731:
3714:
3709:
3708:
3706:
3703:
3702:
3686:
3684:
3681:
3680:
3663:
3658:
3657:
3655:
3652:
3651:
3634:
3629:
3628:
3626:
3623:
3622:
3619:Krylov subspace
3601:
3596:
3595:
3593:
3590:
3589:
3572:
3567:
3566:
3564:
3561:
3560:
3538:
3535:
3534:
3511:
3506:
3505:
3500:
3493:
3492:
3487:
3482:
3479:
3476:
3475:
3458:
3453:
3452:
3450:
3447:
3446:
3445:-orthogonal to
3430:
3428:
3425:
3424:
3407:
3402:
3401:
3399:
3396:
3395:
3372:
3367:
3366:
3359:
3358:
3353:
3348:
3345:
3342:
3341:
3324:
3319:
3318:
3316:
3313:
3312:
3295:
3290:
3289:
3287:
3284:
3283:
3279:
3260:
3259:
3246:
3241:
3240:
3235:
3228:
3227:
3222:
3217:
3215:
3205:
3197:
3196:
3188:
3178:
3177:
3172:
3167:
3165:
3163:
3154:
3150:
3132:
3128:
3119:
3115:
3104:
3101:
3100:
3091:
3087:
3069:
3064:
3063:
3057:
3053:
3044:
3039:
3038:
3025:
3010:
3005:
3004:
2994:
2992:
2989:
2988:
2969:
2965:
2963:
2960:
2959:
2953:
2928:
2924:
2922:
2919:
2918:
2901:
2896:
2895:
2893:
2890:
2889:
2863:
2858:
2857:
2852:
2845:
2844:
2839:
2834:
2832:
2825:
2820:
2819:
2812:
2811:
2806:
2801:
2799:
2797:
2785:
2780:
2779:
2774:
2767:
2766:
2761:
2756:
2754:
2744:
2736:
2735:
2727:
2717:
2716:
2711:
2706:
2704:
2702:
2693:
2689:
2687:
2684:
2683:
2660:
2655:
2654:
2648:
2644:
2635:
2630:
2629:
2614:
2609:
2608:
2606:
2603:
2602:
2579:
2574:
2573:
2564:
2559:
2558:
2553:
2546:
2545:
2540:
2535:
2533:
2526:
2521:
2520:
2515:
2508:
2507:
2502:
2497:
2495:
2493:
2481:
2468:
2463:
2462:
2453:
2448:
2447:
2445:
2442:
2441:
2416:
2411:
2410:
2408:
2405:
2404:
2402:
2374:
2369:
2368:
2366:
2363:
2362:
2346:
2343:
2342:
2325:
2320:
2319:
2317:
2314:
2313:
2290:
2282:
2281:
2273:
2264:
2259:
2258:
2256:
2253:
2252:
2239:
2223:
2212:
2201:
2194:
2179:
2173:
2158:
2130:
2122:
2117:
2106:
2095:
2092:
2091:
2057:
2043:
2029:
2027:
2024:
2023:
1992:
1987:
1986:
1978:
1969:
1962:
1961:
1956:
1955:
1947:
1942:
1935:
1934:
1929:
1928:
1916:
1905:
1897:
1894:
1893:
1884:
1878:
1875:
1869:
1867:
1860:
1840:
1834:
1831:
1825:
1822:
1816:
1792:
1787:
1786:
1784:
1781:
1780:
1763:
1758:
1757:
1755:
1752:
1751:
1748:
1726:
1722:
1720:
1717:
1716:
1700:
1697:
1696:
1686:
1659:
1658:
1654:
1648:
1643:
1642:
1633:
1628:
1627:
1623:
1614:
1605:
1600:
1599:
1595:
1593:
1584:
1580:
1578:
1575:
1574:
1550:
1549:
1538:
1533:
1532:
1523:
1518:
1517:
1516:
1512:
1511:
1505:
1501:
1491:
1490:
1479:
1474:
1473:
1464:
1459:
1458:
1457:
1453:
1452:
1446:
1442:
1436:
1425:
1412:
1407:
1406:
1401:
1394:
1393:
1388:
1383:
1376:
1372:
1366:
1355:
1342:
1337:
1336:
1331:
1324:
1323:
1318:
1313:
1304:
1297:
1296:
1291:
1286:
1283:
1280:
1279:
1258:
1257:
1252:
1247:
1244:
1241:
1240:
1224:
1213:
1211:
1208:
1207:
1184:
1179:
1178:
1173:
1167:
1163:
1157:
1146:
1133:
1128:
1127:
1122:
1113:
1108:
1107:
1101:
1097:
1091:
1080:
1067:
1062:
1061:
1059:
1056:
1055:
1051:in this basis:
1036:
1025:
1023:
1020:
1019:
1002:
997:
996:
994:
991:
990:
973:
968:
967:
965:
962:
961:
941:
938:
937:
915:
912:
911:
888:
883:
882:
877:
870:
869:
864:
859:
856:
853:
852:
836:
834:
831:
830:
814:
811:
810:
787:
782:
781:
766:
761:
760:
749:
746:
745:
741:. Suppose that
726:
724:
721:
720:
704:
702:
699:
698:
682:
680:
677:
676:
660:
658:
655:
654:
629:
624:
616:
602:
595:
594:
589:
588:
580:
566:
558:
553:
540:
539:
535:
530:
522:
511:
506:
499:
498:
493:
492:
490:
487:
486:
464:
462:
459:
458:
433:
428:
421:
420:
415:
414:
412:
409:
408:
389:
387:
384:
383:
356:
350:
328:
323:
322:
320:
317:
316:
300:
298:
295:
294:
270:
268:
265:
264:
225:
223:
220:
219:
197:
194:
193:
177:
175:
172:
171:
170:for the vector
152:
144:
139:
137:
134:
133:
124:
97:Magnus Hestenes
17:
12:
11:
5:
15760:
15750:
15749:
15744:
15727:
15726:
15724:
15723:
15718:
15713:
15708:
15703:
15698:
15692:
15690:
15686:
15685:
15683:
15682:
15677:
15672:
15667:
15662:
15656:
15654:
15650:
15649:
15647:
15646:
15641:
15636:
15626:
15621:
15615:
15613:
15609:
15608:
15606:
15605:
15600:
15598:Floating point
15594:
15592:
15588:
15587:
15580:
15579:
15572:
15565:
15557:
15551:
15550:
15530:
15529:External links
15527:
15526:
15525:
15519:
15506:
15500:
15494:
15471:
15465:
15452:
15446:
15433:
15427:
15408:
15405:
15402:
15401:
15362:
15344:
15321:
15276:
15225:
15211:10.1.1.35.7473
15188:
15174:10.1.1.56.1755
15151:
15140:(1): 220–252.
15120:
15113:
15095:
15068:
15053:
15035:
15028:
15004:
14984:
14935:
14928:
14894:
14867:
14860:
14839:
14814:
14770:
14769:
14767:
14764:
14762:
14761:
14756:
14751:
14746:
14741:
14736:
14731:
14726:
14720:
14713:
14711:
14708:
14695:
14692:
14687:
14683:
14677:
14673:
14669:
14666:
14663:
14660:
14656:
14651:
14647:
14643:
14639:
14635:
14632:
14629:
14624:
14620:
14614:
14610:
14606:
14603:
14600:
14597:
14577:
14574:
14571:
14568:
14548:
14526:
14520:
14517:
14516:
14513:
14510:
14509:
14506:
14503:
14502:
14499:
14496:
14495:
14493:
14488:
14485:
14481:
14476:
14470:
14467:
14464:
14461:
14457:
14452:
14447:
14442:
14437:
14431:
14426:
14423:
14419:
14415:
14412:
14410:
14407:
14403:
14398:
14395:
14388:
14383:
14381:
14378:
14375:
14372:
14368:
14363:
14361:
14353:
14348:
14346:
14343:
14340:
14337:
14333:
14328:
14327:
14318:
14313:
14311:
14308:
14307:
14305:
14300:
14297:
14277:
14274:
14271:
14268:
14265:
14262:
14259:
14247:
14244:
14239:
14236:
14202:
14198:
14194:
14189:
14176:
14173:
14150:preconditioner
14116:
14115:
14043:
14040:
14025:
14021:
13998:
13994:
13973:
13970:
13964:
13961:
13954:
13951:
13948:
13942:
13939:
13912:
13907:
13903:
13899:
13896:
13893:
13890:
13887:
13884:
13879:
13875:
13871:
13868:
13865:
13862:
13859:
13856:
13853:
13850:
13847:
13844:
13824:
13821:
13817:preconditioner
13774:
13771:
13766:
13762:
13749:
13746:
13725:
13711:
13708:
13705:
13700:
13695:
13687:
13681:
13678:
13675:
13670:
13650:
13649:
13633:
13628:
13620:
13614:
13609:
13600:
13597:
13594:
13589:
13581:
13575:
13572:
13569:
13564:
13556:
13551:
13547:
13528:
13527:
13511:
13506:
13498:
13492:
13487:
13479:
13473:
13468:
13463:
13458:
13455:
13452:
13447:
13441:
13434:
13428:
13425:
13422:
13417:
13409:
13404:
13400:
13379:
13376:
13362:
13340:
13318:
13289:
13265:
13259:
13234:
13228:
13205:
13181:
13174:
13168:
13163:
13159:
13136:
13129:
13124:
13118:
13115:
13110:
13105:
13102:
13098:
13074:
13048:
13026:
13004:
12982:
12958:
12953:
12949:
12945:
12922:
12916:
12913:
12908:
12903:
12899:
12875:
12851:
12845:
12842:
12837:
12829:
12824:
12818:
12815:
12810:
12805:
12802:
12798:
12771:
12768:
12763:
12755:
12750:
12744:
12741:
12736:
12731:
12728:
12723:
12720:
12715:
12688:
12682:
12679:
12674:
12669:
12665:
12642:
12638:
12634:
12631:
12603:
12597:
12594:
12589:
12585:
12561:
12538:
12515:
12512:
12507:
12484:
12471:
12468:
12460:preconditioner
12435:
12429:
12426:
12421:
12396:
12391:
12385:
12382:
12377:
12372:
12368:
12362:
12359:
12354:
12330:
12327:
12322:
12308:
12307:
12296:
12292:
12285:
12279:
12274:
12268:
12265:
12258:
12254:
12250:
12244:
12238:
12235:
12219:
12218:
12206:
12200:
12197:
12192:
12187:
12181:
12178:
12169:
12164:
12158:
12155:
12150:
12145:
12141:
12135:
12132:
12127:
12111:
12110:
12104:
12100:The result is
12098:
12093:
12092:
12091:
12079:
12076:
12073:
12070:
12067:
12057:
12044:
12039:
12032:
12028:
12024:
12019:
12016:
12013:
12008:
12003:
11998:
11995:
11992:
11987:
11975:
11959:
11954:
11946:
11940:
11935:
11926:
11923:
11920:
11915:
11907:
11901:
11898:
11895:
11890:
11882:
11877:
11873:
11862:
11849:
11846:
11843:
11838:
11833:
11828:
11825:
11822:
11817:
11811:
11803:
11800:
11797:
11794:
11791:
11780:
11766:
11758:
11745:
11740:
11737:
11730:
11726:
11722:
11717:
11712:
11707:
11702:
11699:
11696:
11691:
11679:
11666:
11661:
11654:
11650:
11646:
11641:
11636:
11631:
11626:
11623:
11620:
11615:
11603:
11587:
11582:
11579:
11571:
11565:
11560:
11551:
11546:
11538:
11532:
11527:
11519:
11514:
11510:
11494:
11482:
11479:
11476:
11466:
11453:
11448:
11443:
11438:
11433:
11421:
11408:
11403:
11398:
11393:
11388:
11382:
11364:
11351:
11346:
11343:
11338:
11334:
11330:
11325:
11320:
11294:
11272:
11266:
11263:
11258:
11234:
11231:
11226:
11207:Preconditioner
11202:
11199:
11181:
11159:
11155:
11151:
11148:
11133:
11130:
11126:Anne Greenbaum
11103:
11099:
11095:
11092:
11088:
11083:
11080:
11077:
11058:
11057:
11046:
11042:
11039:
11036:
11032:
11028:
11025:
11012:
11008:
11004:
11001:
10997:
10992:
10989:
10986:
10980:
10977:
10972:
10968:
10964:
10961:
10954:
10951:
10946:
10942:
10938:
10935:
10908:
10888:
10884:
10880:
10877:
10854:
10851:
10848:
10828:
10825:
10804:
10798:
10795:
10791:
10784:
10778:
10773:
10768:
10763:
10757:
10753:
10750:
10745:
10741:
10737:
10734:
10726:
10723:
10717:
10714:
10687:
10683:
10679:
10676:
10652:
10648:
10644:
10641:
10630:
10629:
10614:
10607:
10601:
10596:
10591:
10586:
10577:
10571:
10567:
10563:
10560:
10555:
10552:
10549:
10543:
10539:
10536:
10533:
10530:
10527:
10525:
10523:
10517:
10511:
10506:
10501:
10496:
10486:
10481:
10475:
10472:
10467:
10463:
10459:
10456:
10449:
10446:
10441:
10437:
10433:
10430:
10422:
10417:
10414:
10411:
10409:
10407:
10401:
10395:
10390:
10385:
10380:
10371:
10367:
10364:
10361:
10358:
10354:
10348:
10344:
10340:
10337:
10334:
10331:
10327:
10318:
10313:
10309:
10305:
10302:
10298:
10294:
10291:
10289:
10287:
10281:
10275:
10269:
10264:
10259:
10255:
10251:
10248:
10244:
10235:
10230:
10226:
10222:
10219:
10215:
10211:
10208:
10206:
10201:
10195:
10190:
10185:
10180:
10175:
10174:
10149:
10144:
10139:
10134:
10129:
10124:
10119:
10114:
10090:
10085:
10061:
10056:
10051:
10046:
10041:
10016:
10003:is the set of
9990:
9986:
9974:
9973:
9962:
9957:
9950:
9947:
9944:
9941:
9938:
9935:
9929:
9921:
9917:
9913:
9910:
9903:
9899:
9894:
9889:
9885:
9869:
9866:
9853:
9849:
9845:
9842:
9822:
9818:
9812:
9809:
9804:
9799:
9796:
9776:
9773:
9770:
9766:
9762:
9758:
9755:
9751:
9746:
9743:
9738:
9716:
9713:
9709:
9705:
9701:
9698:
9673:
9670:
9667:
9664:
9641:
9638:
9635:
9632:
9612:
9592:
9589:
9586:
9583:
9557:
9552:
9526:
9523:
9510:
9503:
9496:
9490:
9489:
9478:
9473:
9467:
9464:
9463:
9460:
9457:
9456:
9454:
9449:
9444:
9438:
9435:
9434:
9431:
9428:
9425:
9424:
9422:
9417:
9414:
9409:
9403:
9400:
9399:
9396:
9393:
9392:
9390:
9385:
9380:
9375:
9368:
9364:
9360:
9355:
9350:
9345:
9340:
9335:
9318:
9311:
9305:
9304:
9293:
9290:
9282:
9276:
9273:
9272:
9269:
9266:
9263:
9262:
9260:
9253:
9247:
9244:
9242:
9239:
9238:
9235:
9232:
9230:
9227:
9226:
9224:
9217:
9211:
9208:
9206:
9203:
9200:
9199:
9197:
9188:
9182:
9179:
9178:
9175:
9172:
9169:
9168:
9166:
9159:
9153:
9150:
9148:
9145:
9142:
9141:
9139:
9131:
9123:
9118:
9115:
9107:
9101:
9096:
9087:
9082:
9074:
9068:
9063:
9055:
9050:
9046:
9029:
9021:
9013:
9006:
9005:
8994:
8989:
8983:
8980:
8979:
8976:
8973:
8970:
8969:
8967:
8962:
8957:
8951:
8948:
8945:
8944:
8941:
8938:
8935:
8934:
8932:
8927:
8924:
8919:
8913:
8910:
8909:
8906:
8903:
8900:
8899:
8897:
8892:
8887:
8882:
8875:
8871:
8867:
8862:
8857:
8852:
8847:
8842:
8825:
8817:
8810:
8809:
8798:
8795:
8787:
8781:
8778:
8775:
8774:
8771:
8768:
8765:
8764:
8762:
8755:
8749:
8746:
8743:
8741:
8738:
8735:
8734:
8732:
8723:
8717:
8714:
8713:
8710:
8707:
8704:
8703:
8701:
8694:
8688:
8685:
8683:
8680:
8677:
8676:
8674:
8666:
8658:
8653:
8645:
8639:
8634:
8625:
8620:
8612:
8606:
8601:
8593:
8588:
8584:
8568:
8560:
8553:
8552:
8541:
8536:
8530:
8527:
8526:
8523:
8520:
8517:
8516:
8514:
8509:
8504:
8498:
8495:
8492:
8491:
8488:
8485:
8482:
8481:
8479:
8472:
8466:
8463:
8461:
8458:
8457:
8454:
8451:
8449:
8446:
8445:
8443:
8436:
8433:
8428:
8423:
8417:
8414:
8411:
8410:
8407:
8404:
8401:
8400:
8398:
8393:
8388:
8383:
8377:
8371:
8367:
8363:
8358:
8353:
8348:
8343:
8338:
8321:
8314:
8308:
8307:
8296:
8291:
8285:
8282:
8281:
8278:
8275:
8274:
8272:
8267:
8262:
8256:
8253:
8250:
8249:
8246:
8243:
8240:
8239:
8237:
8230:
8227:
8222:
8217:
8211:
8208:
8207:
8204:
8201:
8200:
8198:
8193:
8188:
8183:
8176:
8172:
8168:
8163:
8158:
8153:
8148:
8143:
8126:
8120:
8119:
8108:
8105:
8100:
8097:
8092:
8084:
8078:
8075:
8072:
8071:
8068:
8065:
8062:
8061:
8059:
8052:
8046:
8043:
8041:
8038:
8037:
8034:
8031:
8029:
8026:
8025:
8023:
8016:
8010:
8007:
8004:
8002:
7999:
7996:
7995:
7993:
7984:
7978:
7975:
7972:
7971:
7968:
7965:
7962:
7961:
7959:
7952:
7946:
7943:
7940:
7938:
7935:
7932:
7931:
7929:
7921:
7913:
7908:
7905:
7897:
7891:
7886:
7877:
7872:
7864:
7858:
7853:
7845:
7840:
7836:
7819:
7806:
7799:
7792:
7786:
7785:
7774:
7769:
7764:
7759:
7754:
7748:
7745:
7742:
7741:
7738:
7735:
7732:
7731:
7729:
7724:
7719:
7713:
7710:
7709:
7706:
7703:
7702:
7700:
7693:
7687:
7684:
7682:
7679:
7678:
7675:
7672:
7670:
7667:
7666:
7664:
7659:
7654:
7648:
7645:
7644:
7641:
7638:
7637:
7635:
7630:
7625:
7620:
7603:
7592:
7585:
7578:
7572:
7571:
7558:
7552:
7549:
7548:
7545:
7542:
7539:
7538:
7536:
7531:
7526:
7518:
7515:
7510:
7509:
7506:
7501:
7498:
7493:
7492:
7490:
7485:
7481:
7465:
7462:
7458:
7457:
7444:
7438:
7435:
7434:
7431:
7428:
7427:
7425:
7420:
7415:
7410:
7394:
7393:
7382:
7377:
7371:
7368:
7367:
7364:
7361:
7360:
7358:
7353:
7348:
7340:
7336:
7332:
7331:
7326:
7322:
7318:
7317:
7315:
7308:
7302:
7299:
7297:
7294:
7293:
7290:
7287:
7285:
7282:
7281:
7279:
7274:
7270:
7265:
7241:
7238:
7222:new_resid_norm
7216:old_resid_norm
7195:old_resid_norm
7189:new_resid_norm
7150:new_resid_norm
7066:old_resid_norm
7036:old_resid_norm
7012:old_resid_norm
6892:AbstractVector
6880:AbstractVector
6868:AbstractMatrix
6839:
6836:
6830:
6824:
6812:
6806:
6805:
6792:
6787:
6779:
6773:
6768:
6759:
6755:
6751:
6746:
6743:
6738:
6735:
6732:
6727:
6722:
6717:
6712:
6707:
6701:
6695:
6690:
6681:
6677:
6673:
6668:
6663:
6658:
6652:
6645:
6640:
6634:
6631:
6628:
6623:
6616:
6613:
6610:
6606:
6602:
6597:
6592:
6587:
6584:
6579:
6574:
6568:
6561:
6555:
6550:
6523:
6518:
6494:
6491:
6488:
6483:
6470:
6469:
6456:
6453:
6450:
6445:
6437:
6431:
6428:
6425:
6420:
6411:
6407:
6403:
6398:
6395:
6392:
6387:
6384:
6381:
6376:
6371:
6366:
6361:
6356:
6350:
6344:
6341:
6338:
6333:
6324:
6320:
6316:
6311:
6306:
6301:
6295:
6288:
6282:
6279:
6276:
6271:
6253:
6237:
6234:
6229:
6226:
6223:
6218:
6213:
6208:
6203:
6198:
6191:
6187:
6183:
6178:
6173:
6168:
6162:
6148:
6147:
6134:
6129:
6123:
6117:
6113:
6109:
6104:
6099:
6094:
6089:
6086:
6083:
6078:
6062:
6061:
6045:
6040:
6034:
6027:
6021:
6016:
6007:
6002:
5996:
5989:
5983:
5980:
5977:
5972:
5964:
5961:
5956:
5952:
5935:
5919:
5914:
5890:
5887:
5884:
5879:
5864:
5848:
5843:
5836:
5831:
5826:
5821:
5818:
5815:
5810:
5805:
5800:
5797:
5794:
5789:
5776:
5775:
5759:
5754:
5751:
5743:
5737:
5732:
5723:
5718:
5710:
5704:
5699:
5691:
5683:
5678:
5672:
5665:
5659:
5654:
5645:
5640:
5632:
5626:
5621:
5613:
5608:
5604:
5578:
5573:
5549:
5546:
5543:
5538:
5523:
5517:
5514:
5495:
5490:
5487:
5480:
5476:
5472:
5467:
5462:
5457:
5452:
5449:
5446:
5441:
5413:
5410:
5407:
5402:
5399:
5394:
5390:
5386:
5381:
5378:
5375:
5370:
5339:
5336:
5333:
5328:
5325:
5320:
5316:
5312:
5307:
5304:
5301:
5296:
5272:
5268:
5245:
5240:
5237:
5214:
5191:
5188:
5185:
5180:
5177:
5172:
5168:
5164:
5159:
5156:
5153:
5148:
5124:
5119:
5116:
5109:
5105:
5101:
5096:
5091:
5086:
5081:
5078:
5075:
5070:
5046:
5041:
5038:
5033:
5029:
5025:
5020:
5015:
4991:
4986:
4979:
4975:
4971:
4966:
4961:
4956:
4951:
4948:
4945:
4940:
4926:
4923:
4904:
4899:
4871:
4868:
4865:
4860:
4838:
4835:
4830:
4826:
4803:
4798:
4770:
4765:
4751:
4748:
4738:
4733:
4732:
4710:
4707:
4704:
4699:
4689:
4687:
4679:
4677:
4674:
4671:
4668:
4665:
4662:
4658:
4656:
4651:
4646:
4639:
4635:
4631:
4626:
4623:
4620:
4615:
4610:
4605:
4602:
4599:
4594:
4588:
4586:
4578:
4573:
4565:
4559:
4554:
4545:
4542:
4539:
4534:
4526:
4520:
4517:
4514:
4509:
4501:
4496:
4492:
4487:
4485:
4475:
4472:
4469:
4464:
4451:
4449:
4444:
4439:
4436:
4429:
4425:
4421:
4416:
4411:
4406:
4401:
4398:
4395:
4390:
4384:
4382:
4377:
4372:
4365:
4361:
4357:
4352:
4347:
4342:
4337:
4334:
4331:
4326:
4320:
4318:
4310:
4305:
4302:
4294:
4288:
4283:
4274:
4269:
4261:
4255:
4250:
4242:
4237:
4233:
4228:
4226:
4218:
4216:
4213:
4210:
4207:
4204:
4202:
4197:
4192:
4187:
4182:
4177:
4172:
4170:
4160:
4155:
4143:
4138:
4126:
4124:
4119:
4114:
4111:
4106:
4102:
4098:
4093:
4088:
4083:
4081:
4052:
4047:
4024:
4002:
3998:
3994:
3989:
3966:
3961:
3957:
3952:
3949:
3946:
3942:
3938:
3935:
3932:
3929:
3926:
3923:
3920:
3916:
3912:
3909:
3906:
3903:
3900:
3897:
3894:
3891:
3888:
3885:
3882:
3877:
3873:
3869:
3866:
3863:
3860:
3856:
3847:
3842:
3837:
3834:
3829:
3826:
3823:
3820:
3817:
3814:
3809:
3804:
3800:
3779:
3776:
3771:
3766:
3740:
3717:
3712:
3689:
3666:
3661:
3637:
3632:
3617:span the same
3604:
3599:
3575:
3570:
3548:
3545:
3542:
3522:
3519:
3514:
3509:
3503:
3496:
3490:
3485:
3461:
3456:
3433:
3410:
3405:
3383:
3380:
3375:
3370:
3362:
3356:
3351:
3327:
3322:
3298:
3293:
3278:
3275:
3274:
3273:
3258:
3249:
3244:
3238:
3231:
3225:
3220:
3213:
3208:
3203:
3200:
3195:
3191:
3187:
3181:
3175:
3170:
3162:
3157:
3153:
3148:
3144:
3139:
3136:
3131:
3129:
3127:
3122:
3118:
3114:
3110:
3107:
3103:
3102:
3099:
3094:
3090:
3086:
3083:
3080:
3077:
3072:
3067:
3060:
3056:
3052:
3047:
3042:
3037:
3034:
3031:
3028:
3026:
3024:
3019:
3016:
3013:
3008:
3003:
3000:
2997:
2996:
2972:
2968:
2948:
2931:
2927:
2904:
2899:
2886:
2885:
2874:
2866:
2861:
2855:
2848:
2842:
2837:
2828:
2823:
2815:
2809:
2804:
2796:
2788:
2783:
2777:
2770:
2764:
2759:
2752:
2747:
2742:
2739:
2734:
2730:
2726:
2720:
2714:
2709:
2701:
2696:
2692:
2677:
2676:
2663:
2658:
2651:
2647:
2643:
2638:
2633:
2628:
2623:
2620:
2617:
2612:
2596:
2595:
2582:
2577:
2567:
2562:
2556:
2549:
2543:
2538:
2529:
2524:
2518:
2511:
2505:
2500:
2490:
2487:
2484:
2480:
2476:
2471:
2466:
2461:
2456:
2451:
2419:
2414:
2398:
2377:
2372:
2350:
2328:
2323:
2310:
2309:
2298:
2293:
2288:
2285:
2280:
2276:
2272:
2267:
2262:
2235:
2221:
2210:
2199:
2192:
2171:
2156:
2150:
2149:
2138:
2133:
2129:
2125:
2120:
2116:
2113:
2109:
2105:
2102:
2099:
2077:
2076:
2065:
2060:
2056:
2053:
2050:
2046:
2042:
2039:
2036:
2032:
2017:Hessian matrix
2013:
2012:
2001:
1995:
1990:
1985:
1981:
1976:
1972:
1965:
1959:
1954:
1950:
1945:
1938:
1932:
1924:
1921:
1915:
1912:
1908:
1904:
1901:
1882:
1873:
1865:
1858:
1838:
1829:
1820:
1795:
1790:
1766:
1761:
1747:
1744:
1729:
1725:
1704:
1683:
1682:
1671:
1662:
1657:
1651:
1646:
1641:
1636:
1631:
1626:
1621:
1617:
1613:
1608:
1603:
1598:
1592:
1587:
1583:
1568:
1567:
1553:
1547:
1541:
1536:
1531:
1526:
1521:
1515:
1508:
1504:
1500:
1494:
1488:
1482:
1477:
1472:
1467:
1462:
1456:
1449:
1445:
1439:
1434:
1431:
1428:
1424:
1420:
1415:
1410:
1404:
1397:
1391:
1386:
1379:
1375:
1369:
1364:
1361:
1358:
1354:
1350:
1345:
1340:
1334:
1327:
1321:
1316:
1311:
1307:
1300:
1294:
1289:
1261:
1255:
1250:
1227:
1223:
1219:
1216:
1204:
1203:
1192:
1187:
1182:
1176:
1170:
1166:
1160:
1155:
1152:
1149:
1145:
1141:
1136:
1131:
1125:
1121:
1116:
1111:
1104:
1100:
1094:
1089:
1086:
1083:
1079:
1075:
1070:
1065:
1039:
1035:
1031:
1028:
1005:
1000:
976:
971:
945:
925:
922:
919:
899:
896:
891:
886:
880:
873:
867:
862:
839:
818:
807:
806:
795:
790:
785:
780:
777:
774:
769:
764:
759:
756:
753:
729:
707:
685:
663:
651:
650:
639:
636:
632:
627:
623:
619:
615:
612:
609:
605:
598:
592:
587:
583:
579:
576:
573:
569:
565:
561:
556:
552:
549:
543:
538:
533:
529:
525:
521:
518:
514:
509:
502:
496:
467:
455:
454:
443:
440:
436:
431:
424:
418:
392:
366:iteration for
352:Main article:
349:
346:
331:
326:
303:
273:
228:
207:
204:
201:
180:
168:
167:
155:
151:
147:
142:
123:
120:
101:Eduard Stiefel
60:of particular
15:
9:
6:
4:
3:
2:
15759:
15748:
15745:
15743:
15740:
15739:
15737:
15722:
15719:
15717:
15714:
15712:
15709:
15707:
15704:
15702:
15699:
15697:
15694:
15693:
15691:
15687:
15681:
15678:
15676:
15673:
15671:
15668:
15666:
15663:
15661:
15658:
15657:
15655:
15651:
15645:
15642:
15640:
15637:
15634:
15630:
15627:
15625:
15622:
15620:
15617:
15616:
15614:
15610:
15604:
15601:
15599:
15596:
15595:
15593:
15589:
15585:
15578:
15573:
15571:
15566:
15564:
15559:
15558:
15555:
15547:
15543:
15542:
15537:
15533:
15532:
15522:
15516:
15512:
15507:
15505:
15501:
15497:
15491:
15486:
15485:
15478:
15472:
15468:
15462:
15458:
15453:
15449:
15443:
15439:
15434:
15430:
15424:
15419:
15418:
15411:
15410:
15397:
15391:
15377:
15373:
15366:
15355:
15348:
15341:
15337:
15333:
15328:
15326:
15317:
15313:
15309:
15305:
15300:
15295:
15291:
15287:
15280:
15272:
15268:
15263:
15258:
15253:
15248:
15244:
15240:
15236:
15229:
15221:
15217:
15212:
15207:
15203:
15199:
15192:
15184:
15180:
15175:
15170:
15166:
15162:
15155:
15147:
15143:
15139:
15135:
15131:
15124:
15116:
15110:
15106:
15099:
15081:
15080:
15072:
15064:
15060:
15056:
15050:
15046:
15039:
15031:
15025:
15021:
15017:
15016:
15008:
14997:
14996:
14988:
14980:
14976:
14972:
14968:
14963:
14958:
14954:
14950:
14946:
14939:
14931:
14925:
14921:
14917:
14913:
14909:
14903:
14901:
14899:
14890:
14889:
14884:
14883:Polyak, Boris
14878:
14876:
14874:
14872:
14863:
14861:3-540-00217-0
14857:
14853:
14849:
14843:
14834:
14829:
14825:
14818:
14809:
14804:
14800:
14796:
14789:
14785:
14781:
14775:
14771:
14760:
14757:
14755:
14752:
14750:
14747:
14745:
14742:
14740:
14737:
14735:
14732:
14730:
14727:
14724:
14721:
14718:
14715:
14714:
14707:
14693:
14690:
14685:
14675:
14671:
14667:
14664:
14661:
14654:
14649:
14641:
14637:
14633:
14627:
14622:
14612:
14608:
14604:
14601:
14598:
14575:
14572:
14569:
14566:
14546:
14524:
14518:
14511:
14504:
14497:
14491:
14486:
14483:
14479:
14474:
14468:
14465:
14462:
14455:
14440:
14424:
14413:
14408:
14401:
14386:
14379:
14376:
14373:
14366:
14351:
14344:
14341:
14338:
14331:
14316:
14309:
14303:
14298:
14295:
14272:
14269:
14266:
14260:
14257:
14243:
14235:
14233:
14230:for the real
14229:
14225:
14221:
14217:
14196:
14172:
14170:
14166:
14162:
14158:
14153:
14151:
14147:
14143:
14140:
14136:
14132:
14131:sparse matrix
14128:
14124:
14121:
14114:
14111:
14107:
14104:
14101:
14100:
14099:
14097:
14093:
14089:
14085:
14081:
14077:
14074:
14070:
14067:
14063:
14060:
14057:
14053:
14049:
14039:
14023:
14019:
13996:
13992:
13971:
13968:
13962:
13959:
13952:
13949:
13946:
13940:
13937:
13926:
13910:
13905:
13901:
13897:
13891:
13885:
13877:
13873:
13869:
13866:
13860:
13857:
13854:
13848:
13845:
13842:
13834:
13830:
13820:
13818:
13814:
13809:
13803:
13797:
13792:
13788:
13772:
13769:
13764:
13760:
13745:
13743:
13739:
13738:Polak–Ribière
13734:
13732:
13709:
13706:
13703:
13698:
13679:
13676:
13673:
13657:
13655:
13631:
13612:
13598:
13595:
13592:
13573:
13570:
13567:
13554:
13549:
13545:
13537:
13536:
13535:
13533:
13509:
13490:
13477:
13471:
13461:
13456:
13453:
13450:
13439:
13426:
13423:
13420:
13407:
13402:
13398:
13390:
13389:
13388:
13386:
13385:Polak–Ribière
13375:
13306:
13304:
13193:
13161:
13148:
13116:
13113:
13100:
13087:
13063:
12970:
12947:
12934:
12914:
12911:
12901:
12888:
12863:
12843:
12840:
12816:
12813:
12800:
12787:
12784:
12769:
12766:
12742:
12739:
12726:
12721:
12718:
12703:
12700:
12680:
12677:
12667:
12654:
12636:
12620:
12617:
12587:
12574:
12513:
12510:
12467:
12465:
12461:
12456:
12453:
12448:
12427:
12424:
12383:
12380:
12360:
12357:
12328:
12325:
12294:
12272:
12256:
12248:
12224:
12223:
12222:
12198:
12195:
12185:
12156:
12153:
12133:
12130:
12116:
12115:
12114:
12107:
12103:
12099:
12097:
12094:
12077:
12074:
12071:
12068:
12065:
12058:
12042:
12030:
12026:
12022:
12017:
12014:
12011:
12001:
11996:
11993:
11990:
11976:
11957:
11938:
11924:
11921:
11918:
11899:
11896:
11893:
11880:
11875:
11871:
11863:
11847:
11844:
11841:
11831:
11826:
11823:
11820:
11781:
11779:
11775:
11769:
11765:
11762:
11759:
11743:
11728:
11724:
11720:
11715:
11705:
11700:
11697:
11694:
11680:
11664:
11652:
11648:
11644:
11639:
11629:
11624:
11621:
11618:
11604:
11585:
11563:
11549:
11530:
11517:
11512:
11508:
11500:
11499:
11498:
11495:
11480:
11477:
11474:
11467:
11451:
11441:
11436:
11422:
11406:
11396:
11391:
11365:
11349:
11336:
11328:
11323:
11309:
11308:
11307:
11264:
11261:
11232:
11229:
11214:
11208:
11198:
11196:
11179:
11146:
11129:
11127:
11123:
11118:
11090:
11086:
11081:
11078:
11075:
11067:
11063:
11044:
11040:
11037:
11023:
10999:
10995:
10990:
10987:
10984:
10978:
10975:
10959:
10952:
10949:
10933:
10921:
10920:
10919:
10875:
10866:
10852:
10849:
10846:
10826:
10823:
10802:
10796:
10793:
10789:
10771:
10755:
10751:
10748:
10732:
10724:
10721:
10715:
10712:
10703:
10701:
10674:
10666:
10639:
10612:
10594:
10575:
10558:
10553:
10550:
10547:
10541:
10537:
10534:
10531:
10528:
10526:
10504:
10484:
10479:
10473:
10470:
10454:
10447:
10444:
10428:
10420:
10415:
10412:
10410:
10388:
10362:
10356:
10335:
10332:
10329:
10316:
10311:
10303:
10300:
10292:
10290:
10267:
10246:
10233:
10228:
10220:
10217:
10209:
10207:
10188:
10165:
10164:
10163:
10147:
10137:
10132:
10122:
10117:
10088:
10059:
10054:
10049:
10039:
10028:
10014:
10006:
9988:
9960:
9955:
9948:
9945:
9939:
9933:
9927:
9919:
9911:
9908:
9901:
9897:
9892:
9887:
9875:
9874:
9873:
9865:
9864:, see below.
9840:
9810:
9807:
9794:
9774:
9771:
9760:
9744:
9741:
9714:
9711:
9703:
9687:
9668:
9662:
9653:
9636:
9630:
9623:: the larger
9610:
9587:
9581:
9574:
9555:
9540:
9535:
9532:
9522:
9520:
9516:
9509:
9502:
9495:
9476:
9471:
9465:
9458:
9452:
9447:
9442:
9436:
9429:
9426:
9420:
9415:
9412:
9407:
9401:
9394:
9388:
9383:
9378:
9366:
9362:
9358:
9353:
9343:
9338:
9324:
9323:
9322:
9317:
9310:
9291:
9288:
9280:
9274:
9267:
9264:
9258:
9251:
9245:
9240:
9233:
9228:
9222:
9215:
9209:
9204:
9201:
9195:
9186:
9180:
9173:
9170:
9164:
9157:
9151:
9146:
9143:
9137:
9129:
9121:
9099:
9085:
9066:
9053:
9048:
9044:
9036:
9035:
9034:
9028:
9020:
9012:
8992:
8987:
8981:
8974:
8971:
8965:
8960:
8955:
8949:
8946:
8939:
8936:
8930:
8925:
8922:
8917:
8911:
8904:
8901:
8895:
8890:
8885:
8873:
8869:
8865:
8860:
8850:
8845:
8831:
8830:
8829:
8824:
8816:
8796:
8793:
8785:
8779:
8776:
8769:
8766:
8760:
8753:
8747:
8744:
8739:
8736:
8730:
8721:
8715:
8708:
8705:
8699:
8692:
8686:
8681:
8678:
8672:
8664:
8656:
8637:
8623:
8604:
8591:
8586:
8582:
8574:
8573:
8572:
8567:
8559:
8539:
8534:
8528:
8521:
8518:
8512:
8507:
8502:
8496:
8493:
8486:
8483:
8477:
8470:
8464:
8459:
8452:
8447:
8441:
8434:
8431:
8426:
8421:
8415:
8412:
8405:
8402:
8396:
8391:
8386:
8369:
8365:
8361:
8356:
8346:
8341:
8327:
8326:
8325:
8320:
8313:
8294:
8289:
8283:
8276:
8270:
8265:
8260:
8254:
8251:
8244:
8241:
8235:
8228:
8225:
8220:
8215:
8209:
8202:
8196:
8191:
8186:
8174:
8170:
8166:
8161:
8151:
8146:
8132:
8131:
8130:
8125:
8106:
8103:
8098:
8095:
8090:
8082:
8076:
8073:
8066:
8063:
8057:
8050:
8044:
8039:
8032:
8027:
8021:
8014:
8008:
8005:
8000:
7997:
7991:
7982:
7976:
7973:
7966:
7963:
7957:
7950:
7944:
7941:
7936:
7933:
7927:
7919:
7911:
7889:
7875:
7856:
7843:
7838:
7834:
7826:
7825:
7824:
7818:
7812:
7809:
7805:
7798:
7791:
7772:
7767:
7757:
7752:
7746:
7743:
7736:
7733:
7727:
7722:
7717:
7711:
7704:
7698:
7691:
7685:
7680:
7673:
7668:
7662:
7657:
7652:
7646:
7639:
7633:
7628:
7623:
7609:
7608:
7607:
7602:
7598:
7591:
7584:
7577:
7556:
7550:
7540:
7534:
7529:
7524:
7516:
7513:
7499:
7496:
7488:
7483:
7471:
7470:
7469:
7461:
7442:
7436:
7429:
7423:
7418:
7413:
7399:
7398:
7397:
7380:
7375:
7369:
7362:
7356:
7351:
7346:
7338:
7334:
7324:
7320:
7313:
7306:
7300:
7295:
7288:
7283:
7277:
7272:
7255:
7254:
7253:
7251:
7247:
6835:
6829:
6815:
6811:
6790:
6771:
6757:
6753:
6749:
6744:
6736:
6733:
6730:
6720:
6715:
6693:
6679:
6675:
6671:
6666:
6661:
6632:
6629:
6626:
6614:
6611:
6608:
6604:
6600:
6595:
6582:
6577:
6553:
6539:
6538:
6537:
6521:
6492:
6489:
6486:
6454:
6451:
6448:
6429:
6426:
6423:
6409:
6405:
6401:
6396:
6393:
6385:
6382:
6379:
6369:
6364:
6342:
6339:
6336:
6322:
6318:
6314:
6309:
6304:
6280:
6277:
6274:
6260:
6259:
6258:
6248:
6235:
6227:
6224:
6221:
6211:
6206:
6189:
6185:
6181:
6176:
6171:
6151:
6132:
6115:
6111:
6107:
6102:
6092:
6087:
6084:
6081:
6067:
6066:
6065:
6043:
6019:
6005:
5981:
5978:
5975:
5962:
5959:
5954:
5950:
5942:
5941:
5940:
5932:. Initially,
5917:
5888:
5885:
5882:
5846:
5834:
5829:
5824:
5819:
5816:
5813:
5803:
5798:
5795:
5792:
5757:
5735:
5721:
5702:
5689:
5681:
5657:
5643:
5624:
5611:
5606:
5602:
5594:
5593:
5592:
5576:
5547:
5544:
5541:
5513:
5511:
5493:
5478:
5474:
5470:
5465:
5455:
5450:
5447:
5444:
5429:
5411:
5408:
5405:
5392:
5384:
5379:
5376:
5373:
5357:
5355:
5337:
5334:
5331:
5318:
5310:
5305:
5302:
5299:
5270:
5266:
5243:
5189:
5186:
5183:
5170:
5162:
5157:
5154:
5151:
5122:
5107:
5103:
5099:
5094:
5084:
5079:
5076:
5073:
5044:
5031:
5023:
5018:
4989:
4977:
4973:
4969:
4964:
4954:
4949:
4946:
4943:
4929:The formulas
4922:
4920:
4902:
4887:
4869:
4866:
4863:
4836:
4833:
4828:
4824:
4801:
4786:
4768:
4754:We note that
4747:
4745:
4708:
4705:
4702:
4672:
4669:
4666:
4663:
4660:
4649:
4637:
4633:
4629:
4624:
4621:
4618:
4608:
4603:
4600:
4597:
4576:
4557:
4543:
4540:
4537:
4518:
4515:
4512:
4499:
4494:
4490:
4473:
4470:
4467:
4442:
4427:
4423:
4419:
4414:
4404:
4399:
4396:
4393:
4375:
4363:
4359:
4355:
4350:
4340:
4335:
4332:
4329:
4308:
4286:
4272:
4253:
4240:
4235:
4231:
4211:
4208:
4205:
4195:
4185:
4180:
4158:
4141:
4117:
4104:
4096:
4091:
4072:
4071:
4070:
4068:
4050:
3996:
3964:
3959:
3955:
3950:
3947:
3944:
3940:
3936:
3933:
3930:
3927:
3924:
3921:
3918:
3914:
3910:
3907:
3904:
3901:
3898:
3892:
3889:
3886:
3880:
3867:
3864:
3861:
3854:
3845:
3835:
3832:
3807:
3802:
3798:
3777:
3774:
3769:
3753:
3715:
3701:. Therefore,
3664:
3635:
3620:
3602:
3573:
3546:
3543:
3540:
3520:
3517:
3512:
3488:
3459:
3408:
3381:
3378:
3373:
3354:
3325:
3296:
3256:
3247:
3223:
3206:
3193:
3173:
3160:
3155:
3151:
3142:
3137:
3134:
3130:
3120:
3116:
3108:
3105:
3092:
3088:
3081:
3078:
3070:
3058:
3054:
3050:
3045:
3032:
3029:
3027:
3017:
3014:
3011:
2998:
2987:
2986:
2985:
2970:
2966:
2957:
2951:
2947:
2929:
2925:
2902:
2872:
2864:
2840:
2826:
2807:
2794:
2786:
2762:
2745:
2732:
2712:
2699:
2694:
2690:
2682:
2681:
2680:
2661:
2649:
2645:
2641:
2636:
2626:
2621:
2618:
2615:
2601:
2600:
2599:
2580:
2565:
2541:
2527:
2503:
2488:
2485:
2482:
2478:
2474:
2469:
2459:
2454:
2440:
2439:
2438:
2436:
2417:
2401:
2397:
2393:
2375:
2348:
2326:
2296:
2291:
2278:
2270:
2265:
2251:
2250:
2249:
2247:
2243:
2238:
2234:
2229:
2227:
2220:
2217:. Note that
2216:
2209:
2205:
2198:
2191:
2186:
2182:
2177:
2170:
2166:
2162:
2155:
2136:
2127:
2114:
2100:
2090:
2089:
2088:
2086:
2082:
2063:
2054:
2037:
2022:
2021:
2020:
2018:
1999:
1993:
1983:
1974:
1952:
1922:
1919:
1913:
1899:
1892:
1891:
1890:
1889:
1881:
1872:
1864:
1857:
1853:
1849:
1844:
1837:
1828:
1819:
1813:
1811:
1793:
1764:
1743:
1727:
1723:
1702:
1693:
1689:
1669:
1649:
1639:
1634:
1611:
1606:
1590:
1585:
1581:
1573:
1572:
1571:
1545:
1539:
1529:
1524:
1513:
1506:
1502:
1498:
1486:
1480:
1470:
1465:
1454:
1447:
1443:
1437:
1432:
1429:
1426:
1422:
1418:
1413:
1389:
1377:
1373:
1367:
1362:
1359:
1356:
1352:
1348:
1343:
1319:
1309:
1292:
1278:
1277:
1276:
1253:
1221:
1190:
1185:
1168:
1164:
1158:
1153:
1150:
1147:
1143:
1139:
1134:
1114:
1102:
1098:
1092:
1087:
1084:
1081:
1077:
1073:
1068:
1054:
1053:
1052:
1033:
1003:
974:
959:
943:
923:
920:
917:
897:
894:
889:
865:
816:
788:
778:
775:
772:
767:
754:
751:
744:
743:
742:
637:
621:
610:
585:
574:
563:
547:
527:
516:
485:
484:
483:
482:
481:inner product
441:
438:
407:
406:
405:
381:
377:
372:
369:
365:
361:
355:
345:
329:
292:
288:
262:
259:
255:
251:
247:
243:
205:
202:
199:
149:
132:
131:
130:
129:
119:
117:
113:
108:
106:
102:
98:
94:
90:
85:
83:
79:
75:
71:
67:
63:
59:
55:
51:
47:
39:
35:
32:steps, where
31:
26:
21:
15591:Key concepts
15539:
15510:
15483:
15456:
15437:
15416:
15379:. Retrieved
15375:
15365:
15347:
15299:math/0605767
15289:
15285:
15279:
15242:
15238:
15228:
15201:
15197:
15191:
15164:
15160:
15154:
15137:
15133:
15123:
15104:
15098:
15087:. Retrieved
15078:
15071:
15044:
15038:
15014:
15007:
14994:
14987:
14952:
14948:
14938:
14911:
14887:
14851:
14842:
14823:
14817:
14798:
14794:
14774:
14288:, and define
14249:
14241:
14234:everywhere.
14178:
14167:has a large
14164:
14160:
14154:
14145:
14141:
14138:
14126:
14122:
14119:
14117:
14112:
14109:
14105:
14102:
14095:
14091:
14087:
14083:
14075:
14072:
14068:
14065:
14061:
14058:
14051:
14047:
14045:
13826:
13807:
13801:
13795:
13751:
13741:
13735:
13658:
13653:
13651:
13529:
13381:
13307:
13194:
13149:
13088:
12971:
12935:
12889:
12864:
12788:
12785:
12704:
12701:
12655:
12621:
12618:
12575:
12473:
12457:
12451:
12449:
12309:
12220:
12112:
12105:
12101:
12095:
11777:
11773:
11767:
11763:
11760:
11496:
11210:
11135:
11119:
11066:Gauss–Seidel
11059:
10867:
10704:
10698:denotes the
10663:denotes the
10631:
10029:
9975:
9871:
9654:
9536:
9528:
9518:
9514:
9507:
9500:
9493:
9492:The result,
9491:
9315:
9308:
9306:
9026:
9018:
9010:
9007:
8822:
8814:
8811:
8565:
8557:
8554:
8318:
8311:
8309:
8123:
8121:
7816:
7813:
7807:
7803:
7796:
7789:
7787:
7600:
7596:
7589:
7582:
7575:
7573:
7467:
7459:
7395:
7249:
7245:
7243:
6813:
6809:
6807:
6471:
6249:
6152:
6149:
6063:
5777:
5519:
5358:
4928:
4888:method from
4753:
4734:
4692:return
4066:
3754:
3280:
2955:
2949:
2945:
2887:
2678:
2597:
2399:
2395:
2311:
2245:
2236:
2232:
2230:
2224:is also the
2218:
2214:
2207:
2203:
2196:
2189:
2184:
2180:
2175:
2168:
2164:
2160:
2153:
2151:
2084:
2080:
2078:
2014:
1879:
1870:
1862:
1855:
1851:
1847:
1842:
1835:
1826:
1817:
1814:
1809:
1749:
1691:
1687:
1684:
1569:
1205:
809:is a set of
808:
652:
456:
379:
375:
373:
357:
286:
260:
257:
249:
245:
169:
125:
109:
89:optimization
86:
49:
43:
37:
33:
29:
15477:"Chapter 6"
15332:Ross, I. M.
15292:(4): 1267.
15245:: 276–285.
15167:(4): 1305.
13787:line search
10705:This shows
10005:polynomials
9684:is large,
46:mathematics
15736:Categories
15633:algorithms
15381:2023-12-05
15340:1902.09004
15089:2020-03-31
14962:2007.00640
14801:(6): 409.
14766:References
12096:end repeat
11776:exit loop
11205:See also:
9787:such that
4816:. Setting
4682:end repeat
368:eigenvalue
15660:CPU cache
15546:EMS Press
15252:1212.6680
15206:CiteSeerX
15169:CiteSeerX
15063:952572240
14979:0010-3640
14682:‖
14665:−
14659:‖
14619:‖
14602:−
14596:‖
14512:⋮
14425:⋱
14414:⋱
14409:⋱
14261:∈
14232:transpose
14216:Hermitian
14071:, since
14020:γ
13993:γ
13963:˙
13941:˙
13902:γ
13898:−
13883:∇
13874:γ
13870:−
13761:β
13546:β
13462:−
13399:β
13114:−
12912:−
12841:−
12814:−
12767:−
12740:−
12719:−
12678:−
12511:−
12425:−
12381:−
12358:−
12326:−
12267:^
12196:−
12180:^
12154:−
12131:−
12027:β
11872:β
11725:α
11721:−
11649:α
11509:α
11337:−
11262:−
11230:−
11147:κ
11091:κ
11082:−
11076:≈
11038:≫
11024:κ
11000:κ
10991:−
10985:≈
10960:κ
10950:−
10934:κ
10907:∞
10899:tends to
10876:κ
10847:ε
10827:ε
10794:−
10790:ε
10752:
10733:κ
10675:κ
10640:σ
10559:κ
10548:−
10538:
10529:≤
10455:κ
10445:−
10429:κ
10413:≤
10363:λ
10336:σ
10333:∈
10330:λ
10317:∗
10308:Π
10304:∈
10293:≤
10234:∗
10225:Π
10221:∈
10148:∗
10138:−
10089:∗
9985:Π
9916:Π
9912:∈
9893:∗
9884:Π
9841:κ
9808:−
9795:κ
9761:−
9742:−
9704:−
9663:κ
9631:κ
9582:κ
9427:−
9384:≈
9363:α
9265:−
9202:−
9171:−
9144:−
9130:≈
9045:α
8972:−
8947:−
8937:−
8902:−
8891:≈
8870:β
8777:−
8767:−
8745:−
8737:−
8706:−
8679:−
8665:≈
8583:β
8519:−
8508:≈
8494:−
8484:−
8427:−
8413:−
8403:−
8366:α
8362:−
8266:≈
8252:−
8242:−
8171:α
8104:≈
8074:−
8064:−
8006:−
7998:−
7974:−
7964:−
7942:−
7934:−
7835:α
7744:−
7734:−
7658:−
7530:≈
7252:given by
7141:step_size
7114:step_size
7060:step_size
6754:α
6721:−
6676:α
6630:−
6612:−
6605:β
6406:α
6397:−
6370:−
6319:α
6212:−
6186:α
6112:α
6108:−
5963:−
5951:β
5830:β
5825:−
5603:α
5475:α
5471:−
5393:−
5319:−
5267:α
5171:−
5104:α
5100:−
5032:−
4974:α
4825:β
4634:β
4491:β
4424:α
4420:−
4360:α
4232:α
4105:−
3948:−
3934:…
3911:
3905:∈
3890:−
3876:⊤
3865:−
3836:∈
3544:≠
3474:, i.e.
3194:−
3152:α
3147:⇒
3117:α
3089:α
3055:α
2967:α
2926:α
2733:−
2691:α
2646:α
2479:∑
2475:−
2390:, so the
2279:−
2248:th step:
2128:−
2098:∇
1984:∈
1953:−
1794:∗
1724:α
1656:⟩
1625:⟨
1620:⟩
1597:⟨
1582:α
1503:α
1444:α
1423:∑
1374:α
1353:∑
1344:∗
1165:α
1144:∑
1135:∗
1120:⇒
1099:α
1078:∑
1069:∗
1004:∗
921:≠
776:…
635:⟩
614:⟨
608:⟩
578:⟨
572:⟩
551:⟨
537:⟩
520:⟨
330:∗
242:symmetric
203:×
54:algorithm
15689:Software
15653:Hardware
15612:Problems
15513:. SIAM.
15390:cite web
15316:17614913
15271:51978658
14910:(1997).
14885:(1987).
14710:See also
13923:for the
13654:flexible
13534:formula
13387:formula
10839:for any
10777:‖
10762:‖
10665:spectrum
10600:‖
10585:‖
10510:‖
10495:‖
10394:‖
10379:‖
10274:‖
10243:‖
10194:‖
10179:‖
7464:Solution
7180:residual
7162:residual
7135:residual
7129:residual
7024:residual
6964:residual
6928:residual
6853:function
6472:because
4750:Restarts
4456:if
4130:if
3621:. Where
3109:′
2242:residual
2226:residual
1546:⟩
1514:⟨
1487:⟩
1455:⟨
956:forms a
936:. Then
910:for all
56:for the
15548:, 2001
15342:, 2019.
12462:is the
12221:where
9292:0.4122.
8797:0.0088.
3340:, i.e.
2244:at the
2240:be the
2178:equals
1570:and so
1275:yields
851:, i.e.
697:, then
364:Lanczos
360:Arnoldi
289:), and
244:(i.e.,
218:matrix
15711:LAPACK
15701:MATLAB
15517:
15492:
15463:
15444:
15425:
15314:
15269:
15208:
15171:
15111:
15061:
15051:
15026:
14977:
14926:
14858:
14719:(BiCG)
14539:Since
13195:Since
12972:Since
12786:Then:
11806:
11778:end if
11497:repeat
11375:Solve:
11062:Jacobi
10667:, and
10632:where
10580:
10490:
10374:
9976:where
9952:
9931:
9925:
9906:
9537:As an
9466:0.6364
9459:0.0909
9437:0.7229
9430:0.3511
9416:0.4122
9402:0.3384
9395:0.2356
9275:0.7229
9268:0.3511
9210:0.7229
9205:0.3511
9181:0.7492
9174:0.2810
9152:0.7492
9147:0.2810
8982:0.7229
8975:0.3511
8926:0.0088
8912:0.7492
8905:0.2810
8716:0.7492
8709:0.2810
8687:0.7492
8682:0.2810
8529:0.7492
8522:0.2810
8284:0.3384
8277:0.2356
8107:0.2205
7551:0.6364
7541:0.0909
7228:return
6910:eltype
6064:using
5861:. The
5778:since
4221:repeat
4014:where
3790:, then
3533:, for
2679:with
457:Since
293:, and
256:(i.e.
74:sparse
52:is an
48:, the
15696:ATLAS
15357:(PDF)
15336:arXiv
15312:S2CID
15294:arXiv
15267:S2CID
15247:arXiv
15083:(PDF)
14999:(PDF)
14957:arXiv
14791:(PDF)
14725:(CGS)
14129:is a
12702:But:
9727:with
7084:'
7033:while
2954:into
958:basis
404:) if
15675:SIMD
15515:ISBN
15490:ISBN
15461:ISBN
15442:ISBN
15423:ISBN
15396:link
15109:ISBN
15059:OCLC
15049:ISBN
15024:ISBN
14975:ISSN
14924:ISBN
14856:ISBN
14157:LSQR
14096:CGNR
14050:-by-
14011:and
13217:and
12994:and
11774:then
10850:>
10030:Let
9506:and
7156:norm
7039:>
7018:norm
6988:sqrt
6973:norm
6958:copy
6507:and
5137:and
5004:and
3908:span
3588:and
2486:<
2231:Let
960:for
378:and
291:real
110:The
99:and
15665:TLB
15304:doi
15257:doi
15216:doi
15179:doi
15142:doi
15020:195
14967:doi
14916:doi
14828:hdl
14803:doi
14108:=
14098:).
14094:or
14092:CGN
13330:or
11120:No
11064:or
11019:for
10749:log
10535:exp
10326:max
10297:min
10214:min
9655:If
8435:331
8229:331
8099:331
7234:end
7225:end
7042:tol
6994:sum
6904:eps
6898:tol
5939:is
3423:is
2361:at
2163:at
1824:by
1018:of
285:in
252:),
240:is
44:In
15738::
15544:,
15538:,
15480:.
15392:}}
15388:{{
15374:.
15324:^
15310:.
15302:.
15290:29
15288:.
15265:.
15255:.
15243:51
15241:.
15237:.
15214:.
15202:22
15200:.
15177:.
15165:21
15163:.
15136:.
15132:.
15057:.
15022:.
14973:.
14965:.
14953:76
14951:.
14947:.
14922:.
14914:.
14897:^
14870:^
14799:49
14797:.
14793:.
14782:;
14171:.
14137:κ(
14106:Ax
13867::=
13789:,
13770::=
13733:.
13555::=
13408::=
13374:.
13305:.
12887::
12466:.
12447:.
12108:+1
12069::=
12002::=
11881::=
11832::=
11770:+1
11761:if
11706::=
11630::=
11518::=
11478::=
11442::=
11397::=
11329::=
11128:.
11117:.
10865:.
10702:.
10123::=
10027:.
9898::=
9321:.
9033:.
8571:.
8432:73
8226:73
8096:73
7601:Ax
7599:-
7595:=
7517:11
7500:11
7248:=
7246:Ax
7171:@.
7126:@.
7099:@.
7009:))
7003:.^
6919:))
6889:::
6877:::
6865:::
6828:.
5456::=
5385::=
5311::=
5163::=
5085::=
5024::=
4955::=
4921:.
4746:.
4664::=
4609::=
4500::=
4405::=
4341::=
4241::=
4209::=
4186::=
4097::=
3079:=:
2952:+1
2208:Ax
2206:−
2202:=
2183:−
2181:Ax
2167:=
1856:Ax
1854:−
1850:=
1848:Az
1841:=
1742:.
1690:=
1688:Ax
548::=
442:0.
344:.
261:Ax
248:=
105:Z4
15635:)
15631:(
15576:e
15569:t
15562:v
15523:.
15498:.
15469:.
15450:.
15431:.
15398:)
15384:.
15359:.
15338::
15318:.
15306::
15296::
15273:.
15259::
15249::
15222:.
15218::
15185:.
15181::
15148:.
15144::
15138:6
15117:.
15092:.
15065:.
15032:.
15001:.
14981:.
14969::
14959::
14932:.
14918::
14891:.
14864:.
14830::
14811:.
14805::
14694:0
14691:=
14686:2
14676:n
14672:x
14668:W
14662:b
14655:,
14650:k
14646:)
14642:t
14638:/
14634:1
14631:(
14628:=
14623:2
14613:k
14609:x
14605:W
14599:b
14576:b
14573:=
14570:x
14567:W
14547:W
14525:]
14519:0
14505:0
14498:1
14492:[
14487:=
14484:b
14480:,
14475:]
14469:t
14466:+
14463:1
14456:t
14441:t
14402:t
14387:t
14380:t
14377:+
14374:1
14367:t
14352:t
14345:t
14342:+
14339:1
14332:t
14317:t
14310:t
14304:[
14299:=
14296:W
14276:)
14273:1
14270:,
14267:0
14264:(
14258:t
14201:b
14197:=
14193:x
14188:A
14165:A
14161:A
14146:A
14142:A
14139:A
14127:A
14123:A
14120:A
14113:b
14110:A
14103:A
14090:(
14084:A
14076:A
14073:A
14069:b
14066:A
14062:A
14059:A
14052:m
14048:n
14024:b
13997:a
13972:u
13969:=
13960:v
13953:,
13950:v
13947:=
13938:x
13927:,
13911:v
13906:b
13895:)
13892:x
13889:(
13886:f
13878:a
13864:)
13861:v
13858:,
13855:x
13852:(
13849:k
13846:=
13843:u
13835:,
13808:p
13802:z
13796:p
13773:0
13765:k
13726:k
13724:β
13710:,
13707:0
13704:=
13699:k
13694:z
13686:T
13680:1
13677:+
13674:k
13669:r
13632:k
13627:z
13619:T
13613:k
13608:r
13599:1
13596:+
13593:k
13588:z
13580:T
13574:1
13571:+
13568:k
13563:r
13550:k
13510:k
13505:z
13497:T
13491:k
13486:r
13478:)
13472:k
13467:z
13457:1
13454:+
13451:k
13446:z
13440:(
13433:T
13427:1
13424:+
13421:k
13416:r
13403:k
13361:z
13339:L
13317:M
13288:z
13264:T
13258:L
13233:T
13227:L
13204:a
13180:z
13173:T
13167:L
13162:=
13158:a
13135:a
13128:T
13123:)
13117:1
13109:L
13104:(
13101:=
13097:z
13073:a
13047:a
13025:L
13003:L
12981:r
12957:a
12952:L
12948:=
12944:r
12921:r
12915:1
12907:L
12902:=
12898:a
12874:a
12850:r
12844:1
12836:L
12828:T
12823:)
12817:1
12809:L
12804:(
12801:=
12797:z
12770:1
12762:L
12754:T
12749:)
12743:1
12735:L
12730:(
12727:=
12722:1
12714:M
12687:r
12681:1
12673:M
12668:=
12664:z
12641:r
12637:=
12633:z
12630:M
12602:T
12596:L
12593:L
12588:=
12584:M
12560:L
12537:M
12514:1
12506:M
12483:M
12452:M
12434:A
12428:1
12420:M
12395:T
12390:)
12384:1
12376:E
12371:(
12367:A
12361:1
12353:E
12329:1
12321:M
12295:.
12291:x
12284:T
12278:E
12273:=
12264:x
12257:,
12253:M
12249:=
12243:T
12237:E
12234:E
12205:b
12199:1
12191:E
12186:=
12177:x
12168:T
12163:)
12157:1
12149:E
12144:(
12140:A
12134:1
12126:E
12106:k
12102:x
12078:1
12075:+
12072:k
12066:k
12043:k
12038:p
12031:k
12023:+
12018:1
12015:+
12012:k
12007:z
11997:1
11994:+
11991:k
11986:p
11958:k
11953:z
11945:T
11939:k
11934:r
11925:1
11922:+
11919:k
11914:z
11906:T
11900:1
11897:+
11894:k
11889:r
11876:k
11848:1
11845:+
11842:k
11837:r
11827:1
11824:+
11821:k
11816:z
11810:M
11802:e
11799:v
11796:l
11793:o
11790:S
11768:k
11764:r
11744:k
11739:p
11736:A
11729:k
11716:k
11711:r
11701:1
11698:+
11695:k
11690:r
11665:k
11660:p
11653:k
11645:+
11640:k
11635:x
11625:1
11622:+
11619:k
11614:x
11586:k
11581:p
11578:A
11570:T
11564:k
11559:p
11550:k
11545:z
11537:T
11531:k
11526:r
11513:k
11481:0
11475:k
11452:0
11447:z
11437:0
11432:p
11407:0
11402:r
11392:0
11387:z
11381:M
11350:0
11345:x
11342:A
11333:b
11324:0
11319:r
11293:A
11271:A
11265:1
11257:M
11233:1
11225:M
11180:A
11158:)
11154:A
11150:(
11102:)
11098:A
11094:(
11087:2
11079:1
11045:.
11041:1
11035:)
11031:A
11027:(
11011:)
11007:A
11003:(
10996:2
10988:1
10979:1
10976:+
10971:)
10967:A
10963:(
10953:1
10945:)
10941:A
10937:(
10887:)
10883:A
10879:(
10853:0
10824:2
10803:)
10797:1
10783:A
10772:0
10767:e
10756:(
10744:)
10740:A
10736:(
10725:2
10722:1
10716:=
10713:k
10686:)
10682:A
10678:(
10651:)
10647:A
10643:(
10613:,
10606:A
10595:0
10590:e
10576:)
10570:)
10566:A
10562:(
10554:k
10551:2
10542:(
10532:2
10516:A
10505:0
10500:e
10485:k
10480:)
10474:1
10471:+
10466:)
10462:A
10458:(
10448:1
10440:)
10436:A
10432:(
10421:(
10416:2
10400:A
10389:0
10384:e
10370:|
10366:)
10360:(
10357:p
10353:|
10347:)
10343:A
10339:(
10312:k
10301:p
10280:A
10268:0
10263:e
10258:)
10254:A
10250:(
10247:p
10229:k
10218:p
10210:=
10200:A
10189:k
10184:e
10143:x
10133:k
10128:x
10118:k
10113:e
10084:x
10060:k
10055:)
10050:k
10045:x
10040:(
10015:k
9989:k
9961:,
9956:}
9949:1
9946:=
9943:)
9940:0
9937:(
9934:p
9928::
9920:k
9909:p
9902:{
9888:k
9852:)
9848:A
9844:(
9821:)
9817:A
9811:1
9803:M
9798:(
9775:0
9772:=
9769:)
9765:b
9757:x
9754:A
9750:(
9745:1
9737:M
9715:0
9712:=
9708:b
9700:x
9697:A
9672:)
9669:A
9666:(
9640:)
9637:A
9634:(
9611:A
9591:)
9588:A
9585:(
9556:k
9551:x
9519:n
9515:n
9511:0
9508:x
9504:1
9501:x
9497:2
9494:x
9477:.
9472:]
9453:[
9448:=
9443:]
9421:[
9413:+
9408:]
9389:[
9379:1
9374:p
9367:1
9359:+
9354:1
9349:x
9344:=
9339:2
9334:x
9319:1
9316:x
9312:2
9309:x
9289:=
9281:]
9259:[
9252:]
9246:3
9241:1
9234:1
9229:4
9223:[
9216:]
9196:[
9187:]
9165:[
9158:]
9138:[
9122:1
9117:p
9114:A
9106:T
9100:1
9095:p
9086:1
9081:r
9073:T
9067:1
9062:r
9054:=
9049:1
9030:0
9027:α
9022:1
9019:p
9014:1
9011:α
8993:.
8988:]
8966:[
8961:=
8956:]
8950:3
8940:8
8931:[
8923:+
8918:]
8896:[
8886:0
8881:p
8874:0
8866:+
8861:1
8856:r
8851:=
8846:1
8841:p
8826:1
8823:p
8818:0
8815:β
8794:=
8786:]
8780:3
8770:8
8761:[
8754:]
8748:3
8740:8
8731:[
8722:]
8700:[
8693:]
8673:[
8657:0
8652:r
8644:T
8638:0
8633:r
8624:1
8619:r
8611:T
8605:1
8600:r
8592:=
8587:0
8569:1
8566:p
8561:0
8558:β
8540:.
8535:]
8513:[
8503:]
8497:3
8487:8
8478:[
8471:]
8465:3
8460:1
8453:1
8448:4
8442:[
8422:]
8416:3
8406:8
8397:[
8392:=
8387:0
8382:p
8376:A
8370:0
8357:0
8352:r
8347:=
8342:1
8337:r
8322:1
8319:r
8315:1
8312:x
8295:.
8290:]
8271:[
8261:]
8255:3
8245:8
8236:[
8221:+
8216:]
8210:1
8203:2
8197:[
8192:=
8187:0
8182:p
8175:0
8167:+
8162:0
8157:x
8152:=
8147:1
8142:x
8127:1
8124:x
8091:=
8083:]
8077:3
8067:8
8058:[
8051:]
8045:3
8040:1
8033:1
8028:4
8022:[
8015:]
8009:3
8001:8
7992:[
7983:]
7977:3
7967:8
7958:[
7951:]
7945:3
7937:8
7928:[
7920:=
7912:0
7907:p
7904:A
7896:T
7890:0
7885:p
7876:0
7871:r
7863:T
7857:0
7852:r
7844:=
7839:0
7820:0
7817:α
7808:k
7804:p
7800:0
7797:p
7793:0
7790:r
7773:.
7768:0
7763:p
7758:=
7753:]
7747:3
7737:8
7728:[
7723:=
7718:]
7712:1
7705:2
7699:[
7692:]
7686:3
7681:1
7674:1
7669:4
7663:[
7653:]
7647:2
7640:1
7634:[
7629:=
7624:0
7619:r
7604:0
7597:b
7593:0
7590:r
7586:0
7583:x
7579:0
7576:r
7557:]
7535:[
7525:]
7514:7
7497:1
7489:[
7484:=
7480:x
7443:]
7437:1
7430:2
7424:[
7419:=
7414:0
7409:x
7381:,
7376:]
7370:2
7363:1
7357:[
7352:=
7347:]
7339:2
7335:x
7325:1
7321:x
7314:[
7307:]
7301:3
7296:1
7289:1
7284:4
7278:[
7273:=
7269:x
7264:A
7250:b
7231:x
7219:=
7207:*
7204:2
7201:^
7198:)
7192:/
7186:(
7183:+
7177:=
7165:)
7159:(
7153:=
7144:*
7138:-
7132:=
7117:*
7111:+
7108:x
7105:=
7102:x
7093:)
7087:*
7078:(
7075:/
7072:2
7069:^
7063:=
7054:*
7051:A
7048:=
7027:)
7021:(
7015:=
7006:2
7000:x
6997:(
6991:(
6985:=
6982:)
6979:x
6976:(
6967:)
6961:(
6955:=
6946:x
6943:*
6940:A
6937:-
6934:b
6931:=
6922:)
6916:b
6913:(
6907:(
6901:=
6895:;
6886:x
6883:,
6874:b
6871:,
6862:A
6859:(
6825:k
6823:α
6819:β
6814:k
6810:p
6791:k
6786:r
6778:T
6772:k
6767:r
6758:k
6750:1
6745:=
6742:)
6737:1
6734:+
6731:k
6726:r
6716:k
6711:r
6706:(
6700:T
6694:k
6689:r
6680:k
6672:1
6667:=
6662:k
6657:p
6651:A
6644:T
6639:)
6633:1
6627:k
6622:p
6615:1
6609:k
6601:+
6596:k
6591:r
6586:(
6583:=
6578:k
6573:p
6567:A
6560:T
6554:k
6549:p
6522:k
6517:r
6493:1
6490:+
6487:k
6482:r
6455:1
6452:+
6449:k
6444:r
6436:T
6430:1
6427:+
6424:k
6419:r
6410:k
6402:1
6394:=
6391:)
6386:1
6383:+
6380:k
6375:r
6365:k
6360:r
6355:(
6349:T
6343:1
6340:+
6337:k
6332:r
6323:k
6315:1
6310:=
6305:k
6300:p
6294:A
6287:T
6281:1
6278:+
6275:k
6270:r
6254:k
6252:β
6236:,
6233:)
6228:1
6225:+
6222:k
6217:r
6207:k
6202:r
6197:(
6190:k
6182:1
6177:=
6172:k
6167:p
6161:A
6133:k
6128:p
6122:A
6116:k
6103:k
6098:r
6093:=
6088:1
6085:+
6082:k
6077:r
6044:k
6039:p
6033:A
6026:T
6020:k
6015:p
6006:k
6001:p
5995:A
5988:T
5982:1
5979:+
5976:k
5971:r
5960:=
5955:k
5936:k
5934:β
5918:k
5913:p
5889:1
5886:+
5883:k
5878:p
5865:k
5863:β
5847:k
5842:p
5835:k
5820:1
5817:+
5814:k
5809:p
5804:=
5799:1
5796:+
5793:k
5788:r
5758:k
5753:p
5750:A
5742:T
5736:k
5731:p
5722:k
5717:r
5709:T
5703:k
5698:r
5690:=
5682:k
5677:p
5671:A
5664:T
5658:k
5653:r
5644:k
5639:r
5631:T
5625:k
5620:r
5612:=
5607:k
5577:k
5572:r
5548:1
5545:+
5542:k
5537:r
5524:k
5522:α
5494:k
5489:p
5486:A
5479:k
5466:k
5461:r
5451:1
5448:+
5445:k
5440:r
5412:1
5409:+
5406:k
5401:x
5398:A
5389:b
5380:1
5377:+
5374:k
5369:r
5338:1
5335:+
5332:k
5327:x
5324:A
5315:b
5306:1
5303:+
5300:k
5295:r
5271:k
5244:k
5239:p
5236:A
5213:A
5190:1
5187:+
5184:k
5179:x
5176:A
5167:b
5158:1
5155:+
5152:k
5147:r
5123:k
5118:p
5115:A
5108:k
5095:k
5090:r
5080:1
5077:+
5074:k
5069:r
5045:k
5040:x
5037:A
5028:b
5019:k
5014:r
4990:k
4985:p
4978:k
4970:+
4965:k
4960:x
4950:1
4947:+
4944:k
4939:x
4903:k
4898:x
4870:1
4867:+
4864:k
4859:x
4837:0
4834:=
4829:k
4802:0
4797:x
4769:1
4764:x
4739:k
4737:β
4709:1
4706:+
4703:k
4698:x
4673:1
4670:+
4667:k
4661:k
4650:k
4645:p
4638:k
4630:+
4625:1
4622:+
4619:k
4614:r
4604:1
4601:+
4598:k
4593:p
4577:k
4572:r
4564:T
4558:k
4553:r
4544:1
4541:+
4538:k
4533:r
4525:T
4519:1
4516:+
4513:k
4508:r
4495:k
4474:1
4471:+
4468:k
4463:r
4443:k
4438:p
4435:A
4428:k
4415:k
4410:r
4400:1
4397:+
4394:k
4389:r
4376:k
4371:p
4364:k
4356:+
4351:k
4346:x
4336:1
4333:+
4330:k
4325:x
4309:k
4304:p
4301:A
4293:T
4287:k
4282:p
4273:k
4268:r
4260:T
4254:k
4249:r
4236:k
4212:0
4206:k
4196:0
4191:r
4181:0
4176:p
4159:0
4154:x
4142:0
4137:r
4118:0
4113:x
4110:A
4101:b
4092:0
4087:r
4067:0
4051:0
4046:x
4023:A
4001:b
3997:=
3993:x
3988:A
3965:}
3960:}
3956:b
3951:1
3945:k
3941:A
3937:,
3931:,
3928:b
3925:A
3922:,
3919:b
3915:{
3902:y
3899::
3896:)
3893:y
3887:x
3884:(
3881:A
3872:)
3868:y
3862:x
3859:(
3855:{
3846:n
3841:R
3833:y
3828:n
3825:i
3822:m
3819:g
3816:r
3813:a
3808:=
3803:k
3799:x
3778:0
3775:=
3770:0
3765:x
3739:x
3716:k
3711:x
3688:A
3665:i
3660:p
3636:i
3631:r
3603:i
3598:r
3574:i
3569:p
3547:j
3541:i
3521:0
3518:=
3513:j
3508:p
3502:A
3495:T
3489:i
3484:p
3460:j
3455:p
3432:A
3409:i
3404:p
3382:0
3379:=
3374:j
3369:r
3361:T
3355:i
3350:r
3326:j
3321:r
3297:i
3292:r
3257:.
3248:k
3243:p
3237:A
3230:T
3224:k
3219:p
3212:)
3207:k
3202:x
3199:A
3190:b
3186:(
3180:T
3174:k
3169:p
3161:=
3156:k
3143:0
3138:!
3135:=
3126:)
3121:k
3113:(
3106:g
3098:)
3093:k
3085:(
3082:g
3076:)
3071:k
3066:p
3059:k
3051:+
3046:k
3041:x
3036:(
3033:f
3030:=
3023:)
3018:1
3015:+
3012:k
3007:x
3002:(
2999:f
2971:k
2956:f
2950:k
2946:x
2930:k
2903:k
2898:r
2873:,
2865:k
2860:p
2854:A
2847:T
2841:k
2836:p
2827:k
2822:r
2814:T
2808:k
2803:p
2795:=
2787:k
2782:p
2776:A
2769:T
2763:k
2758:p
2751:)
2746:k
2741:x
2738:A
2729:b
2725:(
2719:T
2713:k
2708:p
2700:=
2695:k
2662:k
2657:p
2650:k
2642:+
2637:k
2632:x
2627:=
2622:1
2619:+
2616:k
2611:x
2581:i
2576:p
2566:i
2561:p
2555:A
2548:T
2542:i
2537:p
2528:k
2523:r
2517:A
2510:T
2504:i
2499:p
2489:k
2483:i
2470:k
2465:r
2460:=
2455:k
2450:p
2418:k
2413:p
2400:k
2396:r
2376:k
2371:x
2349:f
2327:k
2322:r
2297:.
2292:k
2287:x
2284:A
2275:b
2271:=
2266:k
2261:r
2246:k
2237:k
2233:r
2222:0
2219:p
2211:0
2204:b
2200:0
2197:p
2193:0
2190:x
2185:b
2176:f
2172:0
2169:x
2165:x
2161:f
2157:0
2154:p
2137:.
2132:b
2124:x
2119:A
2115:=
2112:)
2108:x
2104:(
2101:f
2085:x
2083:(
2081:f
2064:,
2059:A
2055:=
2052:)
2049:)
2045:x
2041:(
2038:f
2035:(
2031:H
2000:.
1994:n
1989:R
1980:x
1975:,
1971:b
1964:T
1958:x
1949:x
1944:A
1937:T
1931:x
1923:2
1920:1
1914:=
1911:)
1907:x
1903:(
1900:f
1883:∗
1880:x
1874:∗
1871:x
1866:0
1863:x
1859:0
1852:b
1843:0
1839:0
1836:x
1830:0
1827:x
1821:∗
1818:x
1810:n
1789:x
1765:k
1760:p
1728:k
1703:n
1692:b
1670:.
1661:A
1650:k
1645:p
1640:,
1635:k
1630:p
1616:b
1612:,
1607:k
1602:p
1591:=
1586:k
1552:A
1540:k
1535:p
1530:,
1525:k
1520:p
1507:k
1499:=
1493:A
1481:i
1476:p
1471:,
1466:k
1461:p
1448:i
1438:n
1433:1
1430:=
1427:i
1419:=
1414:i
1409:p
1403:A
1396:T
1390:k
1385:p
1378:i
1368:n
1363:1
1360:=
1357:i
1349:=
1339:x
1333:A
1326:T
1320:k
1315:p
1310:=
1306:b
1299:T
1293:k
1288:p
1260:T
1254:k
1249:p
1226:b
1222:=
1218:x
1215:A
1191:.
1186:i
1181:p
1175:A
1169:i
1159:n
1154:1
1151:=
1148:i
1140:=
1130:x
1124:A
1115:i
1110:p
1103:i
1093:n
1088:1
1085:=
1082:i
1074:=
1064:x
1038:b
1034:=
1030:x
1027:A
999:x
975:n
970:R
944:P
924:j
918:i
898:0
895:=
890:j
885:p
879:A
872:T
866:i
861:p
838:A
817:n
794:}
789:n
784:p
779:,
773:,
768:1
763:p
758:{
755:=
752:P
728:u
706:v
684:v
662:u
638:.
631:v
626:A
622:,
618:u
611:=
604:v
597:T
591:A
586:,
582:u
575:=
568:v
564:,
560:u
555:A
542:A
532:v
528:,
524:u
517:=
513:v
508:A
501:T
495:u
466:A
439:=
435:v
430:A
423:T
417:u
391:A
380:v
376:u
362:/
325:x
302:b
287:R
272:x
258:x
250:A
246:A
227:A
206:n
200:n
179:x
154:b
150:=
146:x
141:A
38:n
34:n
30:n
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.