Parallel Stochastic Evolution Algorithms for Constrained Multiobjective Optimization

Abstract

Stochastic evolution (StocE) is an evolutionary metaheuristic that has shown to achieve better solution qualities and runtimes when compared to some other well established stochastic metaheuristics. However, unlike these metaheuristics, parallelization of StocE has not been explored before. In this paper, we discuss a comprehensive set of parallel strategies for StocE using a constrained multiobjective VLSI cell placement as an optimization problem. The effectiveness of the proposed strategy is demonstrated by comparing its results with results of parallel SA algorithms on the same optimization problem.