Parallel Meta-heuristic Search

Abstract

We present a general picture of the parallel meta-heuristic search for optimization. We recall the main concepts and strategies in designing parallel meta-heuristics, pointing to a number of contributions that instantiated them for neighborhood-and population-based meta-heuristics, and identify trends and promising research directions. We focus on cooperation-based strategies, which display remarkable performances, in particular on asynchronous cooperation and advanced cooperation strategies that create new information out of exchanged data to enhance the global guidance of the search.