An Implementation of Parallel Computing forHierarchical Logistic Network Design OptimizationUsing PSO

摘要

Recently, we have concerned the strategic optimization on logistic network design anddeveloped an efficient two-level solution method. To cope with extremely large-scaleproblems, in this paper, we propose a novel algorithm for parallel computing. Thereat,noticing the analogy between the two-level algorithms and the master-slaveconfiguration of PC cluster on one hand, and the suitability of the population-basedalgorithm like particle swarm optimization (PSO) on the other hand, we have developeda parallel procedure that can make overhead and idle time extremely small, and bringabout high performance finally.