Parallel Grouping Particle Swarm Optimization with Stream Processing Paradigm
Kun Ma, Shuhui Liu, Yongzheng Lin, Ziqiang Yu, Ke Ji · 2017
Particle swarm optimization (PSO) is a new swarm intelligence technique inspired by social behavior of bird flocking. However, PSO algorithm easily leads convergence difficulty when processing large-scale data. Besides, MapReduce-based PSO need repeating startup of migration strategy job. In this paper, we use a case of live data migration to present parallel grouping PSO with stream processing paradigm. First, fitness, position and velocity are proposed to present particle processing. Second, stream-based grouping are proposed speed up convergence. The experimental results show the best performance of our method in all.