采用概率判定法的分组变异自适应差分进化算法

为了平衡差分进化算法(DE)的全局探索和局部开发过程,提高算法避免陷入局部最优的能力,文中提出采用概率判定法的分组变异自适应差分进化算法(GVADE).GVADE采用概率判定法判定个体进化状态为较好、较差或一般,并根据个体进化状态为个体选择合适的变异算子和控制参数组.同时,为了满足进化状态较差个体变异的需要,设计具有较强全局探索能力的变异算子.在CEC2005标准测试集合上的实验表明,GVADE优于现有的其它DE算法,可以更好地平衡全局探索和局部开发,具有更高的收敛精度.

Group Mutation Adaptive Differential Evolution Algorithm Based on Probability Judgment Method

To balance the global exploration and local development of differential evolution algorithm(DE) and avoid the algorithm falling into the local optimal, a group mutation adaptive differential evolution algorithm based on probability judgment method(GVADE) is proposed. Evolutionary states of an individual are divided into three states based on probability judgment method: better, worse or general. Then, the appropriate mutation operator and control parameter group are applied for the individual. Meanwhile, a mutation operator with strong global exploratory capability is designed to meet the needs of worse evolutionary individual mutation. The experimental results show that GVADE algorithm is superior to the other DE algorithms on the CEC2005 standard testing sets. It can balance the global exploration and local development well with high convergence accuracy.