Fast QAP Solver with ACO and Taboo Search on GPU using Move-Cost Adjusted Thread Assignment
Hannan University, 5-4-33 Amamihigashi, Matsubara, Osaka 580-8502, Japan
GPUs for Genetic and Evolutionary Computation Competition at 2011 Genetic and Evolutionary Computation Conference (GECCO-2011), 2011
@article{tsutsui2011fast,
title={Fast QAP Solver with ACO and Taboo Search on GPU using Move-Cost Adjusted Thread Assignment},
author={Tsutsui, S. and Fujimoto, N.},
year={2011}
}
There are several studies on solving the quadratic assignment problem (QAP) withGPUs using an evolutionary computation. In our previous studies [3], we applied GPU computation to solve quadratic assignment problems (QAPs) using a distributed parallel GA model on GPUs. However, in those studies no local searches were applied. In this QAP solver, we implemented a parallel ACO for QAPs on a GPU by combining tabu search (TS) with ACO in CUDA [4].
November 5, 2011 by hgpu
