/
Lecture  MaxFlow Problem and Augmenting Path Algorithm Lecture  MaxFlow Problem and Augmenting Path Algorithm

Lecture MaxFlow Problem and Augmenting Path Algorithm - PDF document

tatiana-dople
tatiana-dople . @tatiana-dople
Follow
415 views
Uploaded On 2015-05-02

Lecture MaxFlow Problem and Augmenting Path Algorithm - PPT Presentation

The set is the set of nodes in the network The set is the set of directed links ij The set is the set of capacities ij of the links ij The problem is to determine the maximum amount of 64258ow that can be sent from the source node to the sink node T ID: 59062

The set the

Share:

Link:

Embed:

Download Presentation from below link

Download Pdf The PPT/PDF document "Lecture MaxFlow Problem and Augmenting ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.


Presentation Transcript

Lecture20 Outline  Max-Flowproblem  Description  Someexamples  Algorithmforsolvingmax- owproblem  Augmentingpathalgorithm OperationsResearchMethods1 Lecture20 Max-FlowProblem:Multiple-SourcesMultiple-Sinks Wearegivenadirectedcapacitatednetwork(V;E;C)connectingmultiplesourcenodeswithmultiplesinknodes.ThesetVisthesetofnodesandthesetEisthesetofdirectedlinks(i;j)ThesetCisthesetofcapacitiescij0ofthelinks(i;j)2ETheproblemistodeterminethemaximumamountof owthatcanbesentfromthesourcenodestothesinknodes. Thisis Max-FlowProblem formultiple-sourcesandmultiple-sinks OperationsResearchMethods3 Lecture20 Multiple-sourcesmultiple-linksproblemcanbeconvertedtoasingle-sourceandsingle-sinkproblemby  Introducingadummysourcenodethatisconnectedtotheoriginalsourcenodeswithin nitecapacitylinks  Introducingadummysinknodethatisconnectedwiththeoriginalsinknodeswithin nitecapacitylinks OperationsResearchMethods4 Lecture20 Forsmallscaleproblems:Anotheralternativeistointroduceasingle\dummy"sourcenodeconnectedwithalltheoriginalsourcenodesBUT  EachoutgoinglinkfromthedummysourcenodetoanoriginalsourcenodesgetsassignedacapacitythatisequaltothetotalcapacityoftheoutgoinglinksfromsSimilarly,iftherearemultiplesinknodes,weintroduceasingledummysinknodeBUT  Eachincominglinkfromanoriginalsinknodettothedummysinknodegetsassignedacapacitythatisequaltothetotalcapacityoftheincominglinkstosinkt OperationsResearchMethods5 Lecture20 MaximumFlowProblem:MathematicalFormulation WearegivenadirectedcapacitatednetworkG=(V;E;C))withasinglesourceandasinglesinknode.Wewanttoformulatethemax- owproblem.  Foreachlink(i;j)2E,letxijdenotethe owsentonlink(i;j),  Foreachlink(i;j)2E,the owisboundedfromabovebythecapacitycijofthelink:cijxij0  Wehavetospecifythebalanceequations  Allthenodesinthenetworkexceptforthesourceandthesinknodearejust\transit"nodes(in ow=out ow)Xf`j(`;i)2Egx`i�Xfjj(i;j)2Egxij=0foralli6=s;t OperationsResearchMethods7 Lecture20 maximizex12+x13subjecttox12�x25�x24=0balancefornode2x13�x35�x34=0balancefornode3x24+x34�x46=0balancefornode4x25+x35�x56=0balancefornode50x1250x13100x2440x2550x3450x3550x4680x5610Max-FlowisanLPproblem:wecoulduseasimplexmethod OperationsResearchMethods9 Lecture20 Max-FlowAlgorithms Foragivengraph:ndenotesthenumberofnodesmdenotesthenumberofedgesmaxjfjisthemaximumamountof ow Method Complexity Description Simplex { Constrainedbylegal ow Ford-Fulkersonalgorithm O(mmaxjfj) Weightshavetobeintegers Edmonds-Karpalgorithm O(nm2) BasedonFord-Fulkerson Dinitzblocking owalgorithm O(n2m) Buildslayeredgraphs Generalpush-relabelalgorithm O(n2m) Usesapre ow Ford-FulkersonAlgorithmisalsoknownasAugmentingPathalgorithmWewillalsorefertoitasMax-FlowAlgorithm OperationsResearchMethods10 Lecture20 Example:Augmentingpath Supposewechoosetosendthe owalongpath1!3!5!6Supposewechoosetosend4unitsof owalongthispathThentheresidualcapacitiesoflinks(1,2),(3,5)and(5,6)are6,1,and6respectivelyAbetterchoiceistosend5unitsalongthispath.Inthiscase,thecapacityofthelink(3,5)isreached(itsresidualcapacitybecomes0)Wesaythatlink(3,5)issaturated(nomore owcanbesent) OperationsResearchMethods12 Lecture20 ResidualCapacity Thelinksthathavebeenusedtosenda owgetupdatedtore ectthe owpushEverysuchlink(i;j)getsacapacitylabeloftheforma=bwhere  aistheremainingcapacityofthelinkand  bisthetotal owsentalongthatlink  aisviewedasforwardcapacityofthelink  bisviewedasbackwardcapacityofthelink(capacityifwechoosetotraversethelinkintheoppositedirection)  NOTE:a+b=cijThesumofthesenumbersisequaltotheoriginalcapacityofthelink OperationsResearchMethods14