An Exploring Coevolution Multi-Agent System for Multimodal Function Optimization
Xuhua Shi, Haizhen Yu · 2009
Based on evolution theory and multi-agent cooperation mechanism, an Exploring Coevolution Multi-agent System (ECoEMAS) which aims at improving performance of multimodal optimization problems is proposed. Compared with other search methods, ECoEMAS is based on coevolution idea and several novel operations such as dynamic tasks allocation, asynchronous evolution, dynamic solutions memory and hill-valley exploring. These modifications are tested and showed successfully for both static and dynamic benchmarks. Comparative analysis illustrates ECoEMAS's potential value.