Knowledge

N-dimensional polyhedron

Source 📝

956:
can be represented by a system of linear inequalities with rational coefficients, such that the encoding length of each inequality (i.e., the binary encoding length of all rational numbers appearing as coefficients in the inequality) is at most
930:
When solving algorithmic problems on polyhedra, it is important to know whether a certain polyhedron can be represented by an encoding with a small number of bits. There are several measures to the representation complexity of a polyhedron
1577: 1417: 707: 581: 1181: 1221: 192:
A quadrant in the plane. For instance, the region of the cartesian plane consisting of all points above the horizontal axis and to the right of the vertical axis:
1083:, but we do not know the specific inequalities that attain this upper bound. However, it can be proved that the encoding length in any standard representation of 797: 1472: 1600: 1440: 1789: 1714: 1655: 1337: 94:
are scalars. This definition of polyhedra is particularly important as it provides a geometric perspective for problems in
50:. Unlike a 3-dimensional polyhedron, it may be bounded or unbounded. In this terminology, a bounded polyhedron is called a 1613: 246:
A prism of infinite extent. For instance a doubly infinite square prism in 3-space, consisting of a square in the
1227:
Small representation complexity is useful for "rounding" approximate solutions to exact solutions. Specifically:
666: 540: 500:
The representation of a polyhedron by a set of linear inequalities is not unique. It is common to define a
768:
is a general (possibly unbounded) polyhedron, then it can be represented as: P = conv(V) + cone(E), where
1823: 516:
is full-dimensional, then its standard representation is a set of inequalities such that each inequality
59:
Analytically, a convex polyhedron is expressed as the solution set for a system of linear inequalities,
1642:, Graduate Texts in Mathematics, vol. 221 (2nd ed.), New York: Springer-Verlag, p. 26, 1697: 1143: 1186: 481: 1692: 1771: 111: 47: 1763: 1680: 46:, that has flat sides. It may alternatively be defined as the intersection of finitely many 1799: 1724: 1665: 8: 1767: 328: 312: 1759: 737:
is a polytope (a bounded polyhedron), then it can be represented by its set of vertices
1818: 1068:(an upper bound on the facet complexity). This is equivalent to requiring that we know 106:
Many traditional polyhedral forms are n-dimensional polyhedra. Other examples include:
95: 1633: 1785: 1776:, Algorithms and Combinatorics, vol. 2 (2nd ed.), Springer-Verlag, Berlin, 1710: 1651: 1777: 1702: 1643: 1637: 724:. This representation is unique up to the choice of columns of the identity matrix. 1795: 1720: 1661: 627: 286: 282: 39: 786: 1781: 1647: 1079:
In some cases, we know an upper bound on the facet complexity of a polyhedron
1812: 1121:
are contained in the ball of radius 2 around the origin. This means that, if
995:
are finite sets, such that each point in V or E has encoding length at most
35: 746: 595: 316: 32: 781: 212:}. Its sides are the two positive axes, and it is otherwise unbounded. 135: 24: 1691:, Springer Monographs in Mathematics, Dordrecht: Springer, p. 5, 1706: 1243:
is a rational vector whose entries have common denominator at most
918:
are elements of the minimal nonzero faces of the recession cone of
52: 1132:
is a full-dimensional polyhedron with facet complexity at most
1572:{\displaystyle \|qa-c\|+|qb-d|<2^{-3v},0<q<2^{4nv}} 1125:
is a polytope, then P itself is circumscribed in that ball.
728: 537:, and moreover, the inequalities are normalized such that 1183:. Moreover, this ball is contained in the ball of radius 23:
is a geometric object that generalizes the 3-dimensional
1758: 965:+1, since there are n+1 coefficients in each inequality. 31:-dimensional space. It is defined as a set of points in 289:
is a polyhedron. In the Voronoi tessellation of a set
114:
is a polyhedron defined by a single linear inequality,
1475: 1412:{\displaystyle \|qv-w\|<2^{-3f},0<q<2^{4nf}} 1340: 1189: 1146: 669: 594:
is not full-dimensional, then it is contained in its
543: 642:
and a set of inequalities such that each inequality
1773:
Geometric algorithms and combinatorial optimization
1011:These two kinds of complexity are closely related: 1672: 1626: 1571: 1411: 1215: 1175: 812:can be written as a convex combination of at most 701: 575: 832:can be written as a convex combination of points 307:is bounded (hence a traditional polyhedron) when 1810: 1101:implies upper and lower bounds on the volume of 598:, which is defined by a set of linear equations 1450:is a polyhedron with vertex complexity at most 1274:is a polyhedron with vertex complexity at most 925: 1678: 1632: 1314:is a polyhedron with facet complexity at most 1235:is a polyhedron with facet complexity at most 828:-dimensional polyhedron, then every point in 138:is a polyhedron defined by two inequalities, 1491: 1476: 1356: 1341: 780:is another finite set, and cone denotes the 684: 670: 663:, the inequalities are normalized such that 558: 544: 1679:Bruns, Winfried; Gubeladze, Joseph (2009), 1076:(an upper bound on the vertex complexity). 808:-dimensional polytope, then every point in 900:are elements of minimal nonempty faces of 495: 1696: 1469:and q,c,d are integer vectors satisfying 1023:, then P has vertex complexity at most 4 1322:is a rational vector with distance from 1041:, then P has facet complexity at most 3 729:Representation by cones and convex hulls 1811: 1754: 1601:simultaneous diophantine approximation 1441:simultaneous diophantine approximation 772:is (as before) the set of vertices of 1752: 1750: 1748: 1746: 1744: 1742: 1740: 1738: 1736: 1734: 820:. The theorem can be generalized: if 784:. The set cone(E) is also called the 702:{\displaystyle \|a_{i}\|_{\infty }=1} 576:{\displaystyle \|a_{i}\|_{\infty }=1} 343: 1097:The complexity of representation of 816:+1 affinely-independent vertices of 1614:Algorithmic problems on convex sets 409:If a face contains a single point { 13: 1731: 1117:, then trivially, all vertices of 688: 562: 14: 1835: 630:. The standard representation of 215:An octant in Euclidean 3-space, 1599:can be found in polytime using 1439:can be found in polytime using 1334:are integer vectors satisfying 1064:(the number of dimensions) and 1515: 1498: 1297:also satisfies the inequality 1113:has vertex complexity at most 1037:has vertex complexity at most 1: 1619: 1019:has facet complexity at most 720:is orthogonal to each row of 659:defines exactly one facet of 533:defines exactly one facet of 1176:{\displaystyle 2^{-7n^{3}f}} 1140:contains a ball with radius 1007:coefficients in each vector. 926:Complexity of representation 468:has full column rank. Then, 368:(defined by some inequality 7: 1607: 1216:{\displaystyle 2^{5n^{2}f}} 979:can be represented as conv( 606:. It is possible to choose 456:is a polyhedron defined by 101: 10: 1840: 1583:satisfies the inequality c 970:vertex complexity at most 1782:10.1007/978-3-642-78240-4 1648:10.1007/978-1-4613-0019-9 1458:satisfies the inequality 1282:satisfies the inequality 297:corresponding to a point 42:) space of any dimension 1247:, and the distance from 364:if there is a halfspace 172:(which is equivalent to 502:standard representation 496:Standard representation 482:basic feasible solution 397:is the intersection of 250:-plane swept along the 21:-dimensional polyhedron 1685:Polytopes, Rings, and 1573: 1413: 1217: 1177: 798:Carathéodory's theorem 703: 577: 331:of the convex hull of 323:, and otherwise (when 1574: 1414: 1218: 1178: 704: 578: 484:of the linear system 436:-1 dimensional, then 1768:Schrijver, Alexander 1473: 1338: 1187: 1144: 667: 541: 504:for each polyhedron 287:Voronoi tessellation 1824:Linear programming 1569: 1409: 1326:of at most 2, and 1223:around the origin. 1213: 1173: 1003:, since there are 699: 634:contains this set 573: 344:Faces and vertices 96:linear programming 1791:978-3-642-78242-8 1760:Grötschel, Martin 1716:978-0-387-76355-2 1657:978-0-387-00424-2 1419:, then the point 1831: 1803: 1802: 1756: 1729: 1727: 1700: 1681:"Definition 1.1" 1676: 1670: 1668: 1639:Convex Polytopes 1634:Grünbaum, Branko 1630: 1578: 1576: 1575: 1570: 1568: 1567: 1537: 1536: 1518: 1501: 1427:is contained in 1418: 1416: 1415: 1410: 1408: 1407: 1377: 1376: 1222: 1220: 1219: 1214: 1212: 1211: 1207: 1206: 1182: 1180: 1179: 1174: 1172: 1171: 1167: 1166: 944:facet complexity 800:states that, if 708: 706: 705: 700: 692: 691: 682: 681: 582: 580: 579: 574: 566: 565: 556: 555: 432:is nonempty and 352:of a polyhedron 306: 277: 242: 211: 1839: 1838: 1834: 1833: 1832: 1830: 1829: 1828: 1809: 1808: 1807: 1806: 1792: 1757: 1732: 1717: 1707:10.1007/b105283 1698:10.1.1.693.2630 1677: 1673: 1658: 1631: 1627: 1622: 1610: 1595:and the vector 1557: 1553: 1526: 1522: 1514: 1497: 1474: 1471: 1470: 1435:and the vector 1397: 1393: 1366: 1362: 1339: 1336: 1335: 1202: 1198: 1194: 1190: 1188: 1185: 1184: 1162: 1158: 1151: 1147: 1145: 1142: 1141: 928: 916: 910: 898: 892: 872: 866: 859: 852: 844: 838: 731: 718: 687: 683: 677: 673: 668: 665: 664: 657: 647: 628:identity matrix 561: 557: 551: 547: 542: 539: 538: 531: 521: 498: 476:if and only if 472:is a vertex of 383: 373: 346: 298: 255: 216: 193: 187: 177: 170: 160: 153: 143: 129: 119: 104: 92: 83:are vectors in 81: 74: 64: 12: 11: 5: 1837: 1827: 1826: 1821: 1805: 1804: 1790: 1764:Lovász, László 1730: 1715: 1671: 1656: 1624: 1623: 1621: 1618: 1617: 1616: 1609: 1606: 1605: 1604: 1591:. The numbers 1566: 1563: 1560: 1556: 1552: 1549: 1546: 1543: 1540: 1535: 1532: 1529: 1525: 1521: 1517: 1513: 1510: 1507: 1504: 1500: 1496: 1493: 1490: 1487: 1484: 1481: 1478: 1444: 1406: 1403: 1400: 1396: 1392: 1389: 1386: 1383: 1380: 1375: 1372: 1369: 1365: 1361: 1358: 1355: 1352: 1349: 1346: 1343: 1308: 1268: 1263:is in fact in 1225: 1224: 1210: 1205: 1201: 1197: 1193: 1170: 1165: 1161: 1157: 1154: 1150: 1126: 1087:is at most 35 1058:well-described 1050: 1049: 1031: 1009: 1008: 966: 927: 924: 914: 908: 896: 890: 886:+1, such that 870: 864: 857: 850: 842: 836: 787:recession cone 745:is simply the 730: 727: 726: 725: 716: 698: 695: 690: 686: 680: 676: 672: 655: 645: 588: 572: 569: 564: 560: 554: 550: 546: 529: 519: 497: 494: 450: 449: 426: 381: 371: 345: 342: 341: 340: 279: 244: 213: 190: 185: 175: 168: 158: 151: 141: 132: 127: 117: 103: 100: 90: 79: 72: 62: 9: 6: 4: 3: 2: 1836: 1825: 1822: 1820: 1817: 1816: 1814: 1801: 1797: 1793: 1787: 1783: 1779: 1775: 1774: 1769: 1765: 1761: 1755: 1753: 1751: 1749: 1747: 1745: 1743: 1741: 1739: 1737: 1735: 1726: 1722: 1718: 1712: 1708: 1704: 1699: 1694: 1690: 1686: 1682: 1675: 1667: 1663: 1659: 1653: 1649: 1645: 1641: 1640: 1635: 1629: 1625: 1615: 1612: 1611: 1602: 1598: 1594: 1590: 1586: 1582: 1564: 1561: 1558: 1554: 1550: 1547: 1544: 1541: 1538: 1533: 1530: 1527: 1523: 1519: 1511: 1508: 1505: 1502: 1494: 1488: 1485: 1482: 1479: 1468: 1464: 1461: 1457: 1453: 1449: 1445: 1442: 1438: 1434: 1431:. The number 1430: 1426: 1422: 1404: 1401: 1398: 1394: 1390: 1387: 1384: 1381: 1378: 1373: 1370: 1367: 1363: 1359: 1353: 1350: 1347: 1344: 1333: 1329: 1325: 1321: 1317: 1313: 1309: 1307: 1303: 1300: 1296: 1292: 1288: 1285: 1281: 1277: 1273: 1269: 1266: 1262: 1258: 1255:is at most 2/ 1254: 1250: 1246: 1242: 1238: 1234: 1230: 1229: 1228: 1208: 1203: 1199: 1195: 1191: 1168: 1163: 1159: 1155: 1152: 1148: 1139: 1135: 1131: 1127: 1124: 1120: 1116: 1112: 1108: 1107: 1106: 1104: 1100: 1095: 1093: 1090: 1086: 1082: 1077: 1075: 1071: 1067: 1063: 1059: 1055: 1052:A polyhedron 1047: 1044: 1040: 1036: 1032: 1029: 1026: 1022: 1018: 1014: 1013: 1012: 1006: 1002: 998: 994: 990: 986: 982: 978: 974: 973: 967: 964: 960: 955: 951: 950: 945: 941: 938: 937: 936: 934: 923: 921: 917: 907: 903: 899: 889: 885: 881: 877: 873: 863: 856: 849: 845: 835: 831: 827: 823: 819: 815: 811: 807: 803: 799: 795: 793: 789: 788: 783: 779: 775: 771: 767: 762: 760: 756: 752: 748: 744: 740: 736: 723: 719: 712: 696: 693: 678: 674: 662: 658: 651: 648: 641: 637: 633: 629: 625: 621: 617: 613: 609: 605: 601: 597: 593: 589: 586: 570: 567: 552: 548: 536: 532: 525: 522: 515: 511: 510: 509: 507: 503: 493: 491: 487: 483: 479: 475: 471: 467: 463: 459: 455: 447: 443: 439: 435: 431: 427: 424: 420: 416: 412: 408: 407: 406: 404: 400: 396: 392: 388: 384: 377: 374: 367: 363: 359: 355: 351: 339:is unbounded. 338: 334: 330: 326: 322: 318: 314: 310: 305: 301: 296: 292: 288: 284: 280: 275: 271: 268:) : 0 ≤ 267: 263: 259: 253: 249: 245: 240: 236: 232: 228: 224: 220: 214: 209: 205: 201: 197: 191: 188: 181: 178: 171: 164: 161: 154: 147: 144: 137: 133: 130: 123: 120: 113: 109: 108: 107: 99: 97: 93: 86: 82: 75: 68: 65: 57: 55: 54: 49: 45: 41: 37: 34: 30: 26: 22: 20: 1772: 1688: 1684: 1674: 1638: 1628: 1596: 1592: 1588: 1584: 1580: 1466: 1462: 1459: 1455: 1451: 1447: 1436: 1432: 1428: 1424: 1420: 1331: 1327: 1323: 1319: 1315: 1311: 1305: 1301: 1298: 1294: 1290: 1286: 1283: 1279: 1275: 1271: 1264: 1260: 1256: 1252: 1248: 1244: 1240: 1236: 1232: 1226: 1137: 1133: 1129: 1122: 1118: 1114: 1110: 1102: 1098: 1096: 1091: 1088: 1084: 1080: 1078: 1073: 1069: 1065: 1061: 1057: 1053: 1051: 1045: 1042: 1038: 1034: 1027: 1024: 1020: 1016: 1010: 1004: 1000: 999:. Note that 996: 992: 988: 984: 980: 976: 971: 969: 962: 961:. Note that 958: 953: 948: 946: 943: 939: 932: 929: 919: 912: 905: 901: 894: 887: 883: 879: 875: 868: 861: 854: 847: 840: 833: 829: 825: 821: 817: 813: 809: 805: 801: 796: 791: 785: 777: 773: 769: 765: 763: 758: 754: 750: 742: 738: 734: 732: 721: 714: 710: 660: 653: 649: 643: 639: 635: 631: 623: 619: 615: 611: 607: 603: 599: 591: 584: 534: 527: 523: 517: 513: 505: 501: 499: 489: 485: 477: 473: 469: 465: 461: 457: 453: 451: 445: 441: 440:is called a 437: 433: 429: 422: 418: 417:is called a 414: 410: 402: 398: 394: 390: 386: 385:) such that 379: 375: 369: 365: 361: 357: 356:is called a 353: 349: 347: 336: 332: 327:lies on the 324: 320: 311:lies in the 308: 303: 299: 294: 290: 273: 269: 265: 261: 257: 251: 247: 238: 234: 230: 226: 222: 218: 207: 203: 199: 195: 183: 179: 173: 166: 162: 156: 149: 145: 139: 125: 121: 115: 105: 88: 84: 77: 70: 66: 60: 58: 51: 43: 28: 18: 17: 15: 1060:if we know 747:convex hull 713:, and each 596:affine hull 317:convex hull 293:, the cell 48:half-spaces 1813:Categories 1620:References 1056:is called 782:conic hull 622:'), where 610:such that 428:If a face 272:≤ 1, 0 ≤ 136:hyperplane 112:half-space 25:polyhedron 1819:Polyhedra 1693:CiteSeerX 1528:− 1509:− 1492:‖ 1486:− 1477:‖ 1368:− 1357:‖ 1351:− 1342:‖ 1153:− 987:), where 689:∞ 685:‖ 671:‖ 563:∞ 559:‖ 545:‖ 389:contains 348:A subset 229:) : 202:) : 40:Euclidean 1770:(1993), 1636:(2003), 1608:See also 947:at most 709:for all 583:for all 464:, where 452:Suppose 413:}, then 329:boundary 313:interior 254:-axis: 102:Examples 76:, where 53:polytope 1800:1261419 1725:2508056 1689:-theory 1666:1976856 1579:, then 1259:, then 1136:, then 983:)+cone( 824:is any 757:= conv( 315:of the 1798:  1788:  1723:  1713:  1695:  1664:  1654:  1467:b + 2, 1454:, and 1318:, and 1291:b + 2, 1278:, and 1239:, and 968:P has 776:, and 626:is an 419:vertex 36:affine 27:to an 1293:then 1001:v ≥ n 975:, if 963:f ≥ n 952:, if 911:,..., 893:,..., 874:with 860:,..., 839:,..., 804:is a 741:, as 480:is a 442:facet 285:in a 281:Each 237:≥ 0, 233:≥ 0, 206:≥ 0, 1786:ISBN 1711:ISBN 1652:ISBN 1551:< 1545:< 1520:< 1391:< 1385:< 1360:< 1330:and 1072:and 991:and 942:has 904:and 401:and 393:and 358:face 283:cell 276:≤ 1 256:{ ( 241:≥ 0 217:{ ( 210:≥ 0 194:{ ( 182:≤ - 155:and 87:and 38:(or 33:real 1778:doi 1703:doi 1644:doi 1593:q,d 1446:If 1310:If 1270:If 1251:to 1231:If 1128:If 1109:If 1033:If 1015:If 935:: 790:of 764:If 761:). 749:of 733:If 614:= ( 590:If 512:If 508:: 444:of 421:of 405:. 360:of 319:of 98:. 16:An 1815:: 1796:MR 1794:, 1784:, 1766:; 1762:; 1733:^ 1721:MR 1719:, 1709:, 1701:, 1683:, 1662:MR 1660:, 1650:, 1587:≤ 1465:≤ 1306:b. 1304:≤ 1289:≤ 1105:: 1094:. 922:. 882:≤ 846:, 794:. 753:: 652:≤ 638:x= 618:, 602:x= 526:≤ 492:. 488:≤ 486:Ax 460:≤ 458:Ax 378:≤ 335:) 302:∈ 278:}. 264:, 260:, 248:xy 243:}. 225:, 221:, 198:, 189:). 174:-a 165:≥ 148:≤ 134:A 124:≤ 110:A 69:≤ 56:. 1780:: 1728:. 1705:: 1687:K 1669:. 1646:: 1603:. 1597:c 1589:d 1585:x 1581:P 1565:v 1562:n 1559:4 1555:2 1548:q 1542:0 1539:, 1534:v 1531:3 1524:2 1516:| 1512:d 1506:b 1503:q 1499:| 1495:+ 1489:c 1483:a 1480:q 1463:x 1460:a 1456:P 1452:v 1448:P 1443:. 1437:w 1433:q 1429:P 1425:q 1423:/ 1421:w 1405:f 1402:n 1399:4 1395:2 1388:q 1382:0 1379:, 1374:f 1371:3 1364:2 1354:w 1348:v 1345:q 1332:w 1328:q 1324:P 1320:v 1316:f 1312:P 1302:x 1299:a 1295:P 1287:x 1284:a 1280:P 1276:v 1272:P 1267:. 1265:P 1261:y 1257:q 1253:P 1249:y 1245:q 1241:y 1237:f 1233:P 1209:f 1204:2 1200:n 1196:5 1192:2 1169:f 1164:3 1160:n 1156:7 1149:2 1138:P 1134:f 1130:P 1123:P 1119:P 1115:v 1111:P 1103:P 1099:P 1092:f 1089:n 1085:P 1081:P 1074:v 1070:n 1066:f 1062:n 1054:P 1048:. 1046:v 1043:n 1039:v 1035:P 1030:. 1028:f 1025:n 1021:f 1017:P 1005:n 997:v 993:E 989:V 985:E 981:V 977:P 972:v 959:f 954:P 949:f 940:P 933:P 920:P 915:t 913:e 909:1 906:e 902:P 897:s 895:v 891:1 888:v 884:d 880:t 878:+ 876:s 871:t 869:e 867:+ 865:1 862:v 858:1 855:e 853:+ 851:1 848:v 843:s 841:v 837:1 834:v 830:P 826:d 822:P 818:P 814:d 810:P 806:d 802:P 792:P 778:E 774:P 770:V 766:P 759:V 755:P 751:V 743:P 739:V 735:P 722:C 717:i 715:a 711:i 697:1 694:= 679:i 675:a 661:P 656:i 654:b 650:x 646:i 644:a 640:d 636:C 632:P 624:I 620:C 616:I 612:C 608:C 604:d 600:C 592:P 587:. 585:i 571:1 568:= 553:i 549:a 535:P 530:i 528:b 524:x 520:i 518:a 514:P 506:P 490:b 478:v 474:P 470:v 466:A 462:b 454:P 448:. 446:P 438:F 434:n 430:F 425:. 423:P 415:v 411:v 403:P 399:H 395:F 391:P 387:H 382:1 380:b 376:x 372:1 370:a 366:H 362:P 354:P 350:F 337:A 333:S 325:c 321:S 309:c 304:S 300:c 295:A 291:S 274:y 270:x 266:z 262:y 258:x 252:z 239:z 235:y 231:x 227:z 223:y 219:x 208:y 204:x 200:y 196:x 186:1 184:b 180:x 176:1 169:1 167:b 163:x 159:1 157:a 152:1 150:b 146:x 142:1 140:a 131:. 128:1 126:b 122:x 118:1 116:a 91:i 89:b 85:R 80:i 78:a 73:i 71:b 67:x 63:i 61:a 44:n 29:n 19:n

Index

polyhedron
real
affine
Euclidean
half-spaces
polytope
linear programming
half-space
hyperplane
cell
Voronoi tessellation
interior
convex hull
boundary
basic feasible solution
affine hull
identity matrix
convex hull
conic hull
recession cone
Carathéodory's theorem
simultaneous diophantine approximation
simultaneous diophantine approximation
Algorithmic problems on convex sets
Grünbaum, Branko
Convex Polytopes
doi
10.1007/978-1-4613-0019-9
ISBN
978-0-387-00424-2

Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.