Large Neighborhood Local Search Optimization on Graphics Processing Units

International audienceLocal search (LS) algorithms are among the most powerful techniques for solving computationally hard problems in combinatorial optimization. These algorithms could be viewed as ``walks through neighborhoods'' where the walks are performed by iterative procedures that allow to move from a solution to another one in the solution space. In these heuristics, designing operators to explore large promising regions of the search space may improve the quality of the obtained solutions at the expense of a highly computationally process. Therefore, the use of graphics processing units (GPUs) provides an efficient complementary way to speed up the search. However, designing applications on GPU is still complex and error-prone. We provide a methodology to design and implement large neighborhood LS algorithms on GPU. Finding efficient mappings of the neighborhood structures onto the GPU threads organization is a challenging issue dealt with in this paper. The work has been experimented for binary problems by deploying multiple neighborhood structures. The obtained results are convincing both in terms of efficiency, quality and robustness of the provided solutions at run time