摘要

Considering that the basic Shuffled Complex Evolution(SCE) algorithm has the drawbacks of slow rate of convergence and poor quality of solutions, an individual evolution strategy based shuffled complex evolution (IES-SCE) algorithm is proposed in this paper. In the IES-SCE algorithm, the strategy of evolution is changed to along the best point in the population. This strategy improves the speed of finding optimal solution and the quality of solution. Meantime the global convergence and evolutionary process of IES-SCE are analyzed with the theory of stochastic processes, the Markov chain is established for IES-SCE algorithm and its global convergence is also proved. The experimental results on six benchmark functions demonstrate its global convergence.

全文