A Novel Evolutionary Algorithm for Function Optimization Using MEC

This paper proposes a novel evolutionary algorithm that integrates Mind Evolutionary Computation (MEC) and non-uniform mutation. The algorithm greatly extends MEC to explore the tradeoff between exploration and exploitation for optimizing multimodal functions. Similartaxis mecha- nism drives the proposed algorithm to locate multiple local optima, while non-uniform method locates the global area cooperatively. Moreover, the 1/5 rule is adopted to guide the search direction based on information obtained from feed- back. The proposed algorithm is experimentally testified with a test suits containing six complex multimodal func- tion optimization problems. All experiments demonstrate that the proposed algorithm is competitive with other evo- lutionary algorithms published to date in both convergence velocity and solution quality.