(1)(State Key Lab of Software Engineering,Wuhan University,Wuhan 430072,China)(2)(Computer School,Wuhan University,Wuhan 430072,China)3(School of Information Science and Technology,Jiujiang University,Jiujiang 332005,China)
Abstract:In the field of evolutionary computation,particle swarm optimization algorithm has a certain number of advantages,such as easy realization,fast convergence and less tuning parameters.However,with the size of the problem increasing,particle swarm optimization algorithm gets stuck in the dilemma of low accuracy and time-consuming,thus this paper presents a distributed particle swarm optimization based on resilient distributed datasets.In this algorithm,the global population is divided into several sub-populations,namely islands and uses data structure RDD to make each of islands correspond to one partition.With the partition computing model of RDD,the algorithm has achieved a parallel particle swarm optimization in distributed platform.Finally,this algorithm has been compared with some modified PSO on 11 benchmark functions including unimodal functions and multimodal functions,the result demonstrates that this algorithm has high precision as well as obvious acceleration.