U. Aldasoro Marcellan, L. F. Escudero Bueno, M. Merino Maestre, G. Pérez Sainz de Rozas

The aim of this paper is firstly to help to understand the paradigm of parallel programming for solving a sequence of mixed integer optimization problems and secondly to present some detailed explanations about its practical implementation. We have developed a C++ experimental code that uses the IBM ILOG CPLEX optimizer within the COmputational INfrastructure for Operations Research (COIN-OR) and the Message Passing Interface (MPI) library for the distributed memory environment. The practical experiments have been performed at the ARINA computational cluster provided by the SGI/IZO-SGIker at the UPV/EHU and report the optimization of 44 asymmetric problems by using different parallelization strategies and number of threads. Also, we extend this analysis to the resolution of a testbed of multistage stochastic mixed 0-1 instances that have been used as a pilot case for testing our exact decomposition algorithm Branch and Fix Coordination Multistage (BFC-MS).

Palabras clave: optimization message passing interface parallel computing COIN-OR CPLEX

A. R. N. Simbo, L. F. Escudero Bueno, D. Ríos Insua

P. Olaso Redondo, A. Alonso-Ayuso, L. F. Escudero Bueno, C. Pizarro Romero