Knowledge

Circulation problem

Source 📝

1272: 535: 31:
also being required for the source and sink (i.e. there are no special nodes). In variants of the problem, there are multiple commodities flowing through the network, and a cost on the flow.
898: 733: 379: 821: 1021: 438: 1832: 1609: 1168: 1340: 593: 1521: 1922: 1873: 1738: 1697: 1650: 1439: 1477: 264: 187: 110: 72: 1770: 1122: 1095: 1048: 296: 1547: 1398: 1378: 1292: 1068: 655: 635: 615: 227: 207: 150: 130: 1345: 1526: 2013: 1173: 453: 2008: 951: 909: 836: 664: 307: 1351: 755: 545:
In a multi-commodity circulation problem, you also need to keep track of the flow of the individual commodities:
957: 1969: 386: 444:
Finding a flow assignment satisfying the constraints gives a solution to the given circulation problem.
1878: 1655: 1775: 1552: 1357: 935:
Below are given some problems, and how to solve them with the general circulation setup given above.
1127: 1297: 555: 924: 1482: 1884: 1837: 1702: 1661: 1614: 1403: 1444: 1981: 1950: 234: 157: 80: 42: 1743: 1100: 1073: 1026: 269: 920: 1532: 1383: 1363: 1277: 1053: 640: 620: 600: 212: 192: 135: 115: 2002: 939:
Minimum cost multi-commodity circulation problem - Using all constraints given above.
908:
For the circulation problem, many polynomial algorithms have been developed (e.g.,
24: 912:, 1972; Tarjan 1987-1988). Tardos found the first strongly polynomial algorithm. 916: 1941:Éva Tardos (1985). "A strongly polynomial minimum cost circulation algorithm". 830:
The conservation constraint must be upheld individually for the commodities:
27:
problems, with the added constraint of a lower bound on edge flows, and with
1989: 1954: 1985: 1970:"On the complexity of timetable and multi-commodity flow problems" 1881:- Let all capacities be unlimited, and find a flow of 1 for 945:
Multi-commodity circulation - Solve without optimising cost.
919:
for integer flows. For fractional flows, it is solvable in
948:
Simple circulation - Just use one commodity, and no cost.
942:
Minimum cost circulation problem - Use a single commodity
1267:{\displaystyle l_{i}(t_{i},s_{i})=u(t_{i},s_{i})=d_{i}} 745:
There is also a lower bound on each flow of commodity.
1887: 1840: 1778: 1746: 1705: 1664: 1617: 1555: 1535: 1485: 1447: 1406: 1386: 1366: 1300: 1280: 1176: 1130: 1103: 1076: 1056: 1029: 960: 915:
For the case of multiple commodities, the problem is
839: 758: 667: 643: 623: 603: 558: 530:{\displaystyle \sum _{(v,w)\in E}c(v,w)\cdot f(v,w).} 456: 447:
In the minimum cost variant of the problem, minimize
389: 310: 272: 237: 215: 195: 160: 138: 118: 83: 45: 1360:- Set all costs to 0, and add an edge from the sink 1916: 1867: 1826: 1764: 1732: 1691: 1644: 1603: 1541: 1515: 1471: 1433: 1392: 1372: 1334: 1286: 1266: 1162: 1116: 1089: 1062: 1042: 1015: 892: 815: 727: 649: 629: 609: 587: 529: 432: 373: 290: 258: 221: 201: 181: 144: 124: 104: 66: 893:{\displaystyle \ \sum _{w\in V}f_{i}(u,w)=0.} 8: 1740:for all edges in the graph, and add an edge 728:{\displaystyle \,f(v,w)=\sum _{i}f_{i}(v,w)} 374:{\displaystyle l(v,w)\leq f(v,w)\leq u(v,w)} 816:{\displaystyle \,l_{i}(v,w)\leq f_{i}(v,w)} 440:(flow cannot appear or disappear in nodes). 1968:S. Even and A. Itai and A. Shamir (1976). 1906: 1886: 1839: 1777: 1745: 1704: 1663: 1616: 1554: 1534: 1484: 1446: 1405: 1385: 1365: 1346:Minimum cost multi-commodity flow problem 1305: 1299: 1279: 1258: 1242: 1229: 1207: 1194: 1181: 1175: 1151: 1138: 1129: 1108: 1102: 1081: 1075: 1055: 1034: 1028: 1004: 991: 978: 965: 959: 863: 847: 838: 792: 764: 759: 757: 704: 694: 668: 666: 642: 622: 602: 564: 559: 557: 461: 455: 394: 388: 309: 271: 236: 214: 194: 159: 137: 117: 82: 44: 23:and its variants are a generalisation of 1924:commodities, one for each pair of nodes. 1016:{\displaystyle K_{i}(s_{i},t_{i},d_{i})} 923:, as one can formulate the problem as a 1933: 16:Generalization of network flow problems 433:{\displaystyle \sum _{w\in V}f(u,w)=0} 1529:- First find the maximum flow amount 7: 1348:- As above, but minimize the cost. 14: 1527:Minimum cost maximum flow problem 189:, upper bound on flow from node 112:, lower bound on flow from node 1827:{\displaystyle l(t,s)=u(t,s)=1} 1604:{\displaystyle l(t,s)=u(t,s)=m} 1903: 1891: 1856: 1844: 1815: 1803: 1794: 1782: 1759: 1747: 1721: 1709: 1680: 1668: 1633: 1621: 1592: 1580: 1571: 1559: 1501: 1489: 1463: 1451: 1422: 1410: 1323: 1311: 1248: 1222: 1213: 1187: 1157: 1131: 1010: 971: 881: 869: 810: 798: 782: 770: 722: 710: 684: 672: 582: 570: 521: 509: 500: 488: 474: 462: 421: 409: 368: 356: 347: 335: 326: 314: 285: 273: 253: 241: 176: 164: 99: 87: 61: 49: 1: 1354:- As above, with 1 commodity. 1163:{\displaystyle (t_{i},s_{i})} 1335:{\displaystyle l_{i}(u,v)=0} 588:{\displaystyle \,f_{i}(v,w)} 266:, cost of a unit of flow on 1656:Single-source shortest path 541:Multi-commodity circulation 2030: 1974:SIAM Journal on Computing 1516:{\displaystyle c(t,s)=-1} 1352:Minimum cost flow problem 1917:{\displaystyle v(v-1)/2} 1868:{\displaystyle c(t,s)=0} 1733:{\displaystyle c(u,v)=1} 1692:{\displaystyle l(u,v)=0} 1645:{\displaystyle c(t,s)=0} 1434:{\displaystyle l(t,s)=0} 1879:All-pairs shortest path 1472:{\displaystyle u(t,s)=} 1918: 1869: 1828: 1766: 1734: 1693: 1646: 1605: 1543: 1517: 1473: 1435: 1394: 1374: 1336: 1288: 1268: 1164: 1118: 1091: 1064: 1044: 1017: 910:Edmonds–Karp algorithm 894: 817: 729: 651: 631: 611: 597:The flow of commodity 589: 531: 434: 375: 292: 260: 259:{\displaystyle c(v,w)} 223: 203: 183: 182:{\displaystyle u(v,w)} 146: 126: 106: 105:{\displaystyle l(v,w)} 68: 67:{\displaystyle G(V,E)} 2014:Mathematical problems 1919: 1870: 1829: 1767: 1765:{\displaystyle (t,s)} 1735: 1694: 1647: 1606: 1544: 1518: 1474: 1436: 1395: 1375: 1337: 1289: 1269: 1165: 1119: 1117:{\displaystyle t_{i}} 1092: 1090:{\displaystyle s_{i}} 1065: 1045: 1043:{\displaystyle d_{i}} 1018: 895: 818: 730: 652: 632: 612: 590: 532: 435: 376: 301:and the constraints: 293: 291:{\displaystyle (v,w)} 261: 224: 204: 184: 147: 127: 107: 69: 2009:Network flow problem 1980:(4). SIAM: 691–703. 1885: 1838: 1776: 1744: 1703: 1662: 1615: 1553: 1533: 1483: 1445: 1404: 1384: 1364: 1358:Maximum flow problem 1342:for all other edges. 1298: 1278: 1274:for all commodities 1174: 1128: 1101: 1074: 1054: 1027: 1023:denotes a demand of 958: 952:Multi-commodity flow 837: 756: 665: 641: 621: 601: 556: 454: 387: 308: 270: 235: 213: 193: 158: 136: 116: 81: 43: 39:Given flow network 21:circulation problem 1955:10.1007/BF02579369 1914: 1865: 1824: 1762: 1730: 1689: 1642: 1601: 1549:. Then solve with 1539: 1513: 1469: 1431: 1390: 1370: 1332: 1284: 1264: 1160: 1114: 1087: 1060: 1040: 1013: 890: 858: 813: 725: 699: 647: 627: 607: 585: 527: 484: 430: 405: 371: 288: 256: 219: 199: 179: 142: 122: 102: 64: 1542:{\displaystyle m} 1393:{\displaystyle s} 1373:{\displaystyle t} 1287:{\displaystyle i} 1124:, create an edge 1063:{\displaystyle i} 843: 842: 826: 825: 741: 740: 690: 650:{\displaystyle w} 630:{\displaystyle v} 610:{\displaystyle i} 457: 390: 222:{\displaystyle w} 202:{\displaystyle v} 145:{\displaystyle w} 125:{\displaystyle v} 29:flow conservation 2021: 1994: 1993: 1988:. Archived from 1965: 1959: 1958: 1938: 1923: 1921: 1920: 1915: 1910: 1874: 1872: 1871: 1866: 1833: 1831: 1830: 1825: 1771: 1769: 1768: 1763: 1739: 1737: 1736: 1731: 1698: 1696: 1695: 1690: 1651: 1649: 1648: 1643: 1610: 1608: 1607: 1602: 1548: 1546: 1545: 1540: 1522: 1520: 1519: 1514: 1478: 1476: 1475: 1470: 1440: 1438: 1437: 1432: 1399: 1397: 1396: 1391: 1379: 1377: 1376: 1371: 1341: 1339: 1338: 1333: 1310: 1309: 1293: 1291: 1290: 1285: 1273: 1271: 1270: 1265: 1263: 1262: 1247: 1246: 1234: 1233: 1212: 1211: 1199: 1198: 1186: 1185: 1169: 1167: 1166: 1161: 1156: 1155: 1143: 1142: 1123: 1121: 1120: 1115: 1113: 1112: 1096: 1094: 1093: 1088: 1086: 1085: 1069: 1067: 1066: 1061: 1049: 1047: 1046: 1041: 1039: 1038: 1022: 1020: 1019: 1014: 1009: 1008: 996: 995: 983: 982: 970: 969: 931:Related problems 899: 897: 896: 891: 868: 867: 857: 840: 822: 820: 819: 814: 797: 796: 769: 768: 750: 749: 737:The total flow. 734: 732: 731: 726: 709: 708: 698: 656: 654: 653: 648: 636: 634: 633: 628: 616: 614: 613: 608: 594: 592: 591: 586: 569: 568: 550: 549: 536: 534: 533: 528: 483: 439: 437: 436: 431: 404: 380: 378: 377: 372: 297: 295: 294: 289: 265: 263: 262: 257: 228: 226: 225: 220: 208: 206: 205: 200: 188: 186: 185: 180: 151: 149: 148: 143: 131: 129: 128: 123: 111: 109: 108: 103: 73: 71: 70: 65: 2029: 2028: 2024: 2023: 2022: 2020: 2019: 2018: 1999: 1998: 1997: 1986:10.1137/0205048 1967: 1966: 1962: 1940: 1939: 1935: 1931: 1883: 1882: 1836: 1835: 1774: 1773: 1742: 1741: 1701: 1700: 1660: 1659: 1613: 1612: 1551: 1550: 1531: 1530: 1481: 1480: 1443: 1442: 1402: 1401: 1382: 1381: 1362: 1361: 1301: 1296: 1295: 1276: 1275: 1254: 1238: 1225: 1203: 1190: 1177: 1172: 1171: 1147: 1134: 1126: 1125: 1104: 1099: 1098: 1077: 1072: 1071: 1052: 1051: 1030: 1025: 1024: 1000: 987: 974: 961: 956: 955: 933: 921:polynomial time 906: 859: 835: 834: 788: 760: 754: 753: 700: 663: 662: 639: 638: 619: 618: 599: 598: 560: 554: 553: 543: 452: 451: 385: 384: 306: 305: 268: 267: 233: 232: 211: 210: 191: 190: 156: 155: 134: 133: 114: 113: 79: 78: 41: 40: 37: 17: 12: 11: 5: 2027: 2025: 2017: 2016: 2011: 2001: 2000: 1996: 1995: 1992:on 2013-01-12. 1960: 1949:(3): 247–255. 1932: 1930: 1927: 1926: 1925: 1913: 1909: 1905: 1902: 1899: 1896: 1893: 1890: 1876: 1864: 1861: 1858: 1855: 1852: 1849: 1846: 1843: 1823: 1820: 1817: 1814: 1811: 1808: 1805: 1802: 1799: 1796: 1793: 1790: 1787: 1784: 1781: 1761: 1758: 1755: 1752: 1749: 1729: 1726: 1723: 1720: 1717: 1714: 1711: 1708: 1688: 1685: 1682: 1679: 1676: 1673: 1670: 1667: 1653: 1641: 1638: 1635: 1632: 1629: 1626: 1623: 1620: 1600: 1597: 1594: 1591: 1588: 1585: 1582: 1579: 1576: 1573: 1570: 1567: 1564: 1561: 1558: 1538: 1524: 1512: 1509: 1506: 1503: 1500: 1497: 1494: 1491: 1488: 1468: 1465: 1462: 1459: 1456: 1453: 1450: 1430: 1427: 1424: 1421: 1418: 1415: 1412: 1409: 1389: 1380:to the source 1369: 1355: 1349: 1343: 1331: 1328: 1325: 1322: 1319: 1316: 1313: 1308: 1304: 1283: 1261: 1257: 1253: 1250: 1245: 1241: 1237: 1232: 1228: 1224: 1221: 1218: 1215: 1210: 1206: 1202: 1197: 1193: 1189: 1184: 1180: 1159: 1154: 1150: 1146: 1141: 1137: 1133: 1111: 1107: 1084: 1080: 1059: 1050:for commodity 1037: 1033: 1012: 1007: 1003: 999: 994: 990: 986: 981: 977: 973: 968: 964: 949: 946: 943: 940: 932: 929: 925:linear program 905: 902: 901: 900: 889: 886: 883: 880: 877: 874: 871: 866: 862: 856: 853: 850: 846: 828: 827: 824: 823: 812: 809: 806: 803: 800: 795: 791: 787: 784: 781: 778: 775: 772: 767: 763: 743: 742: 739: 738: 735: 724: 721: 718: 715: 712: 707: 703: 697: 693: 689: 686: 683: 680: 677: 674: 671: 659: 658: 646: 626: 606: 595: 584: 581: 578: 575: 572: 567: 563: 542: 539: 538: 537: 526: 523: 520: 517: 514: 511: 508: 505: 502: 499: 496: 493: 490: 487: 482: 479: 476: 473: 470: 467: 464: 460: 442: 441: 429: 426: 423: 420: 417: 414: 411: 408: 403: 400: 397: 393: 382: 370: 367: 364: 361: 358: 355: 352: 349: 346: 343: 340: 337: 334: 331: 328: 325: 322: 319: 316: 313: 299: 298: 287: 284: 281: 278: 275: 255: 252: 249: 246: 243: 240: 230: 218: 198: 178: 175: 172: 169: 166: 163: 153: 141: 121: 101: 98: 95: 92: 89: 86: 63: 60: 57: 54: 51: 48: 36: 33: 15: 13: 10: 9: 6: 4: 3: 2: 2026: 2015: 2012: 2010: 2007: 2006: 2004: 1991: 1987: 1983: 1979: 1975: 1971: 1964: 1961: 1956: 1952: 1948: 1944: 1943:Combinatorica 1937: 1934: 1928: 1911: 1907: 1900: 1897: 1894: 1888: 1880: 1877: 1862: 1859: 1853: 1850: 1847: 1841: 1821: 1818: 1812: 1809: 1806: 1800: 1797: 1791: 1788: 1785: 1779: 1756: 1753: 1750: 1727: 1724: 1718: 1715: 1712: 1706: 1686: 1683: 1677: 1674: 1671: 1665: 1657: 1654: 1639: 1636: 1630: 1627: 1624: 1618: 1598: 1595: 1589: 1586: 1583: 1577: 1574: 1568: 1565: 1562: 1556: 1536: 1528: 1525: 1510: 1507: 1504: 1498: 1495: 1492: 1486: 1466: 1460: 1457: 1454: 1448: 1428: 1425: 1419: 1416: 1413: 1407: 1387: 1367: 1359: 1356: 1353: 1350: 1347: 1344: 1329: 1326: 1320: 1317: 1314: 1306: 1302: 1281: 1259: 1255: 1251: 1243: 1239: 1235: 1230: 1226: 1219: 1216: 1208: 1204: 1200: 1195: 1191: 1182: 1178: 1152: 1148: 1144: 1139: 1135: 1109: 1105: 1082: 1078: 1057: 1035: 1031: 1005: 1001: 997: 992: 988: 984: 979: 975: 966: 962: 953: 950: 947: 944: 941: 938: 937: 936: 930: 928: 926: 922: 918: 913: 911: 903: 887: 884: 878: 875: 872: 864: 860: 854: 851: 848: 844: 833: 832: 831: 807: 804: 801: 793: 789: 785: 779: 776: 773: 765: 761: 752: 751: 748: 747: 746: 736: 719: 716: 713: 705: 701: 695: 691: 687: 681: 678: 675: 669: 661: 660: 644: 624: 604: 596: 579: 576: 573: 565: 561: 552: 551: 548: 547: 546: 540: 524: 518: 515: 512: 506: 503: 497: 494: 491: 485: 480: 477: 471: 468: 465: 458: 450: 449: 448: 445: 427: 424: 418: 415: 412: 406: 401: 398: 395: 391: 383: 365: 362: 359: 353: 350: 344: 341: 338: 332: 329: 323: 320: 317: 311: 304: 303: 302: 282: 279: 276: 250: 247: 244: 238: 231: 216: 196: 173: 170: 167: 161: 154: 139: 119: 96: 93: 90: 84: 77: 76: 75: 58: 55: 52: 46: 34: 32: 30: 26: 22: 1990:the original 1977: 1973: 1963: 1946: 1942: 1936: 934: 914: 907: 829: 744: 544: 446: 443: 300: 38: 28: 25:network flow 20: 18: 917:NP-complete 2003:Categories 1929:References 35:Definition 1898:− 1508:− 852:∈ 845:∑ 786:≤ 692:∑ 504:⋅ 478:∈ 459:∑ 399:∈ 392:∑ 351:≤ 330:≤ 904:Solution 209:to node 132:to node 1658:- Let 1479:∞ and 1294:. Let 841:  74:with: 1772:with 1400:with 1170:with 1070:from 954:- If 617:from 1834:and 1699:and 1611:and 19:The 1982:doi 1951:doi 1097:to 637:to 2005:: 1976:. 1972:. 1945:. 1441:, 927:. 888:0. 657:. 1984:: 1978:5 1957:. 1953:: 1947:5 1912:2 1908:/ 1904:) 1901:1 1895:v 1892:( 1889:v 1875:. 1863:0 1860:= 1857:) 1854:s 1851:, 1848:t 1845:( 1842:c 1822:1 1819:= 1816:) 1813:s 1810:, 1807:t 1804:( 1801:u 1798:= 1795:) 1792:s 1789:, 1786:t 1783:( 1780:l 1760:) 1757:s 1754:, 1751:t 1748:( 1728:1 1725:= 1722:) 1719:v 1716:, 1713:u 1710:( 1707:c 1687:0 1684:= 1681:) 1678:v 1675:, 1672:u 1669:( 1666:l 1652:. 1640:0 1637:= 1634:) 1631:s 1628:, 1625:t 1622:( 1619:c 1599:m 1596:= 1593:) 1590:s 1587:, 1584:t 1581:( 1578:u 1575:= 1572:) 1569:s 1566:, 1563:t 1560:( 1557:l 1537:m 1523:. 1511:1 1505:= 1502:) 1499:s 1496:, 1493:t 1490:( 1487:c 1467:= 1464:) 1461:s 1458:, 1455:t 1452:( 1449:u 1429:0 1426:= 1423:) 1420:s 1417:, 1414:t 1411:( 1408:l 1388:s 1368:t 1330:0 1327:= 1324:) 1321:v 1318:, 1315:u 1312:( 1307:i 1303:l 1282:i 1260:i 1256:d 1252:= 1249:) 1244:i 1240:s 1236:, 1231:i 1227:t 1223:( 1220:u 1217:= 1214:) 1209:i 1205:s 1201:, 1196:i 1192:t 1188:( 1183:i 1179:l 1158:) 1153:i 1149:s 1145:, 1140:i 1136:t 1132:( 1110:i 1106:t 1083:i 1079:s 1058:i 1036:i 1032:d 1011:) 1006:i 1002:d 998:, 993:i 989:t 985:, 980:i 976:s 972:( 967:i 963:K 885:= 882:) 879:w 876:, 873:u 870:( 865:i 861:f 855:V 849:w 811:) 808:w 805:, 802:v 799:( 794:i 790:f 783:) 780:w 777:, 774:v 771:( 766:i 762:l 723:) 720:w 717:, 714:v 711:( 706:i 702:f 696:i 688:= 685:) 682:w 679:, 676:v 673:( 670:f 645:w 625:v 605:i 583:) 580:w 577:, 574:v 571:( 566:i 562:f 525:. 522:) 519:w 516:, 513:v 510:( 507:f 501:) 498:w 495:, 492:v 489:( 486:c 481:E 475:) 472:w 469:, 466:v 463:( 428:0 425:= 422:) 419:w 416:, 413:u 410:( 407:f 402:V 396:w 381:, 369:) 366:w 363:, 360:v 357:( 354:u 348:) 345:w 342:, 339:v 336:( 333:f 327:) 324:w 321:, 318:v 315:( 312:l 286:) 283:w 280:, 277:v 274:( 254:) 251:w 248:, 245:v 242:( 239:c 229:, 217:w 197:v 177:) 174:w 171:, 168:v 165:( 162:u 152:, 140:w 120:v 100:) 97:w 94:, 91:v 88:( 85:l 62:) 59:E 56:, 53:V 50:( 47:G

Index

network flow
Edmonds–Karp algorithm
NP-complete
polynomial time
linear program
Multi-commodity flow
Minimum cost multi-commodity flow problem
Minimum cost flow problem
Maximum flow problem
Minimum cost maximum flow problem
Single-source shortest path
All-pairs shortest path
doi
10.1007/BF02579369
"On the complexity of timetable and multi-commodity flow problems"
doi
10.1137/0205048
the original
Categories
Network flow problem
Mathematical problems

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