Generating cluster submodels from a multistage stochastic mixed integer optimization model using break stage
Data(s) |
22/07/2013
22/07/2013
01/07/2013
|
---|---|
Resumo |
We present a scheme to generate clusters submodels with stage ordering from a (symmetric or a nonsymmetric one) multistage stochastic mixed integer optimization model using break stage. We consider a stochastic model in compact representation and MPS format with a known scenario tree. The cluster submodels are built by storing first the 0-1 the variables, stage by stage, and then the continuous ones, also stage by stage. A C++ experimental code has been implemented for reordering the stochastic model as well as the cluster decomposition after the relaxation of the non-anticipativiy constraints until the so-called breakstage. The computational experience shows better performance of the stage ordering in terms of elapsed time in a randomly generated testbed of multistage stochastic mixed integer problems. |
Identificador |
1134-8984 |
Idioma(s) |
eng |
Relação |
Biltoki;2013.02 |
Direitos |
info:eu-repo/semantics/openAccess |
Palavras-Chave | #stochastic optimization #scenario cluster partitioning #break stage #C++ #MPS |
Tipo |
info:eu-repo/semantics/workingPaper |