Parallel genetic algorithm for the uncapacited single allocation hub location problem on GPU

Abstract

A parallel genetic algorithm (GA) implemented on GPU clusters is proposed to solve the Uncapacitated Single Allocation Hub Location problem. The GA uses binary and integer encoding with genetic operators adapted to this problem. Our GA is improved by initially locating hubs at middle nodes. In our implementation we use the power of the GPU to compute in parallel several initial solutions, varying the number of hubs. The obtained experimental results compared with the best known solutions on all benchmarks. They show that our approach outperforms most well-known heuristics in terms of solution quality and time execution. Also it allowed to solve instances problem unsolved before.