基于非均匀变异的进化算法对高维多峰函数的收敛性分析

对基于非均匀变异算子的进化算法的实验和机理分析已经证明了该算法模型的良好特性,最近基于非均匀变异算子的进化算法模型求解一维多峰函数问题的收敛性已经得到证明.基于马尔科夫过程理论,对基于非均匀变异算子的一般性进化算法模型和一般性高维多峰函数的收敛性给出证明,并基于典型算例与同类典型算法进行性能比较,数值试验表明算法模型具有很好的性能表现和应用前景.

CONVERGENCE ANALYSIS OF EVOLUTIONARY ALGORITHM BASED ON NON-UNIFORM MUTATION FOR MULTI-DIMENSIONAL AND MULTIMODAL FUNCTIONS

The empirical and mechanism analysis based on evolutionary algorithm(EA)for non-uniform mutation operator show its excellent performance. Recently the convergence analysis of EA with non-uniform mutation operator is provided for one-dimensional functions. In this paper, based on the Markov process theory the convergence is generalized to general multi-dimensional and multimodal functions. The performance comparison with FEP indicates that the proposed EA model has good performance and potential application.