Abstract
In this paper, we study various parallelization schemes for the Variable Neighborhood Search (VNS) metaheuristic on a CPU-GPU system via OpenMP and OpenACC. A hybrid parallel VNS method is applied to recent benchmark problem instances for the multi-product dynamic lot sizing problem with product returns and recovery, which appears in reverse logistics and is known to be NP-hard. We report our findings regarding these parallelization approaches and present promising computational results.
Figure: A hybrid CPU-GPU parallelization scheme
Citation
Antoniadis N. and Sifaleras A., ”A hybrid CPU-GPU parallelization scheme of variable neighborhood search for inventory optimization problems”, Electronic Notes in Discrete Mathematics, Elsevier B.V., Vol. 58, pp. 47-54, 2017.
@article{ANTONIADIS201747,
title = {A hybrid CPU-GPU parallelization scheme of variable neighborhood search for inventory optimization problems},
journal = {Electronic Notes in Discrete Mathematics},
volume = {58},
pages = {47-54},
year = {2017},
note = {4th International Conference on Variable Neighborhood Search},
issn = {1571-0653},
doi = {https://doi.org/10.1016/j.endm.2017.03.007},
url = {https://www.sciencedirect.com/science/article/pii/S1571065317300434},
author = {Nikolaos Antoniadis and Angelo Sifaleras},
keywords = {Variable Neighborhood Search, Parallel Computing, CPU-GPU computing, OpenMP, OpenACC},
abstract = {In this paper, we study various parallelization schemes for the Variable Neighborhood Search (VNS) metaheuristic on a CPU-GPU system via OpenMP and OpenACC. A hybrid parallel VNS method is applied to recent benchmark problem instances for the multi-product dynamic lot sizing problem with product returns and recovery, which appears in reverse logistics and is known to be NP-hard. We report our findings regarding these parallelization approaches and present promising computational results.}
}