P. Olaso Redondo, L. F. Escudero Bueno, C. Pizarro Romero, A. Alonso-Ayuso
We present an inexact algorithm for solving huge instances of multistage mixed integer optimization problems under uncertainty, represented by multistage scenarios trees. The algorithm is based on a metaheuristic methodology since in an iterative way repeatedly solves at each node of the tree up to optimality an stochastic multistage mixed integer model for the subtree whose root-node is the given node related to the scenario group in the original scenario tree. This multistage submodel to solve will relax in each optimization the non-anticipativity constraints related to the successor nodes of the nodes in the subtree that the modeler selects for being considered without any relaxation of the constraints in the problem.
Palabras clave: stochastic mixed integer, optimization, heuristics, metaheuristics, parallel concurrent programming
Programado
VD5 Optimización bajo incertidumbre
20 de abril de 2012 15:30
Sala Viena