一种基于混沌序列的改进自适应差分进化算法

摘要

Differential evolution（DE） is a fast optimization algorithm with high reliability and good robustness. It is widely applied in many fields with good results. A modified algorithm is proposed in this paper to improve the searching performance of DE. The algorithm improves the performance of global optimization by using logistic chaotic sequence to initialize population, re- places fixed parameters of DE by adaptive probability and, at the same time, and introduces sub- population local search strategy. The modified algorithm is validated for several benchmark func- tions. Experimental results demonstrate that the optimal solutions obtained by using the proposed algorithm are better than those obtained by other DE algorithms.%差分进化（DE）算法被认为是一种可靠、鲁棒性好且快速的优化算法，在许多领域得到广泛的应用，并取得了良好效果。为提高DE的寻优性能，提出了一种改进的自适应DE算法。改进后的算法采用Logistic混沌序列产生初始种群，用自适应的变异、交叉参数代替标准DE固定参数，并引入子种群内部搜索策略，使算法具有较好的全局搜索能力。采用经典测试函数对算法进行验证，结果表明，改进后的算法寻优精度和收敛速度得到了有效的提高，具有较好的实用性。