针对差分进化算法在复杂优化问题求解时后期收敛速度慢、易陷入局部最优和参数设置繁琐等问题,提出一种基于新变异策略的动态自适应差分进化算法p-ADE.首先,新变异策略中通过利用种群的全局最优解和目标个体的历史最优解引导种群搜索方向,为下一代个体的生成引入更多有效的方向性信息,避免差分向量中个体随机选择导致的搜索盲目性.其次,为加快收敛速度、提高算法稳定性、避免参数设置的繁琐与不精确,提出一种参数动态自适应调整策略,动态平衡算法局部搜索与全局搜索间的关系,有效调节个体在进化过程中的变异程度.在10个Benchmark函数上的实验结果表明,p-ADE相对于多种先进DE优化策略和全局优化算法在收敛精度、速度和鲁棒性上均具有明显优势.%To overcome the slow convergence speed, premature convergence and tedious parameter settings of the differential evolution when solving complex optimization problems, a dynamic adaptive differential evolution, called p-ADE, based on a novel mutation strategy, is proposed. Firstly, the best global solution and the best previous solution of each individual are utilized in the new mutation strategy to guide the search direction by introducing more effective directional information, avoiding the search blindness brought by the random selection of individuals in the difference vector. Secondly, a self-adaptive parameter setting strategy is designed, which is utilized to balance the global and local search dynamically. Experimental results on 10 benchmark functions show that p-ADE can effectively improve the global search ability of DE and outperforms several state-of-the-art optimization algorithms in terms of the main performance indexes.
展开▼
