A New Particle Swarm Optimization Based on Differential Dimension Coefficient

摘要：

The particle swarm optimization (PSO) algorithm has exhibited good performance on well-known numerical test problems. But PSO suffers from premature convergence on multimodal test problems. J.Rige introduce Diversity-Guided Particle Swarm Optimizer (ARPSO) to improve the performance on multimodal function. As is analyzed below, test functions gradient and dimension have an important impact on the diversity value in ARPSO evolution formula. We introduce a new PSO called Differential Dimension Guided Particle Swarm Optimization (DDPSO) to improve the performance of PSO on the high order multimodal function. Here we add two coefficients to the evolution formula of ARPSO, One is differential coefficient called dif which is proportioned to the function order. The other is dimension coefficient called dim. Then we take four benchmark multimodal functions as test function and make two experiments. Results show that DDPSO outperform ARPSO on high order multimodal function especially when the population size is small.

关键词： Particle Swarm Optimization Attractive andRepulsive Particle Swarm Optimization Differential Dimension Guided Particle Swarm Optimization

作者: Zhihuang Liu Yimin Yang

作者单位: Applied Mathematics School of Guangdong University of Technology,Guangzhou Higher Education Mega Center,Panyu District,P,R China

会议类型: 国际会议

会议名称: 2008高等智能国际会议(2008 International Conference on Advanced Intelligence)

会议地点: 北京

会议语种:英文

在线出版日期: 2008-10-18（万方平台首次上网日期，不代表论文的发表时间）