Ruttningen av trafik i IP-nätverk sker ofta med hjälp av bågvikter som bestämmer vilken väg trafiken tar (kortastevägruttning). Problemet här är att avgöra ifall det existerar en uppsättning vikter givet ett önskat ruttningsschema. Den hör rapporten undersöker prestandan hos ett antal modeller och optimeringsprogram avsedda att lösa denna typ av problem som ofta kallas inversa kortastevägruttningsproblemet. Undersökningen visar att det existerar en stor skillnad mellan modellerna och optimeringsprogrammen och att modellen baserad på cykelbaser löst med CPLEXdualopt lösaren är snabbast. / The routing of traffic in IP networks is often done with a set of weights that determinewhich way the traffic will go (shortest path routing). The problem here is todetermine if there exists a set of weights for a desired routing pattern. This thesis willinvestigate the performance of a number of different models and solvers for solvingthis type of problem which is usually called the inverse shortest path routing (ISPR)problem. The models tested are the same as described in [1]. The different solversused are mainly the linear CPLEX solvers but also a few multi commodity networksolvers. The tests showed that there is a big performance difference between the models andsolvers and that the cycle bases model solved with the CPLEX dualopt solver wasthe fastest overall.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:liu-59536 |
| Date | January 2010 |
| Creators | Sandberg, Richard |
| Publisher | Linköpings universitet, Optimeringslära |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0034 seconds