【摘要】 差分进化算法(DE)是一种著名的处理非线性复杂问题的优化技术,为改进其计算开销大,参数设置与问题本身特性过于相关等缺陷,提出了一种新的差分进化算法(MDE)。它结合了反向学习(OBL)和自适应控制机制来进行参数调整,从而能加速算法收敛,同时提高求解成功率。在Matlab环境中进行的测试实验结果表明,MDE在收敛速度、鲁棒性和计算开销等方面的性能在大部分测试用例上优于已有的多种算法,证明混合策略是一种行之有效的差分进化算法的研究路径。更多还原

【Abstract】 Differential evolution(DE)is a well-known optimization technique to deal with nonlinear and complex problems.In order to tackle the problems,such as much overhead,problem-dependent parameters,etc. This paper presents a mixed DE algorithm,called MDE,by employing opposition-based learning(OBL)and a self-adapting mechanism to adjust parameters to improve the convergence and robustness. Experiments in Matlab show that the proposed approach MDE outperforms many existing algorithms on convergence,robustness and overhead,proving that hybrid is an effective path on DE research.更多还原