改进自适应微分进化算法求解全局优化问题  

Improved adaptive differential evolution algorithm for global optimization

作　　者：王世豪[1,2] 杨红雨[1,2] 李玉贞[3] 刘洪[2] 杨波[2] Wang Shihao Yang Hongyu Li Yuzhen Liu Hong Yang Bo(School of Aeronautics ＆ Astronautics National Key Laboratory of Air Traffic Control Automatien System Technology, Sichuan University, Chengdu 610065, China Shanghai Electrical Apparatus Research Institute, Shanghai 200063, China)

机构地区：[1]四川大学空天科学与工程学院,成都610065 [2]四川大学国家空管自动化系统技术重点实验室,成都610065 [3]上海电器科学研究所,上海200063

出　　处：《计算机应用研究》2016年第12期3634-3637,共4页Application Research of Computers

基　　金：国家"863"计划资助项目(2013AA013802);国家空管科研资助项目(GKG201403004)

摘　　要：针对微分进化(differential evolution,DE)算法在进化后期收敛速度慢、收敛精度低、易陷入局部最优解等缺点。通过改进DE的变异方程,并引入一种新的控制参数自适应策略,提出了一种改进自适应微分进化(improved adaptive differential evolution,IADE)算法。进化过程中IADE将根据个体适应值与父代平均适应值之间的关系动态地调整控制参数。同时,采用10个常用于优化算法比较的标准函数对IADE和其他改进DE算法进行对比实验。实验结果表明,IADE算法不仅能够显著地提高收敛速度和收敛精度,而且具有非常好的鲁棒性,从而使得该算法能够满足过程优化的实时性、准确性以及稳定性要求。Differential evolution （DE） algorithm has some disadvantages, such as slow convergence speed, low convergence precision and easy to fall into local optimal solution in the early stages of the evolution. The paper proposed an improved adaptive differential evolution （IADE） algorithm by improving the mutation equation of DE and introducing a new control parameters adaption strategy. In the process of the evolution, the control parameters will be dynamically adjusted by comparing individ- ual fitness with average fitness of the parent population. Meanwhile, the paper chose the ten standard functions commonly used for the comparison of optimization algorithm to perform the comparative test of IADE and the other improved DE algorithms, and the experimental results show that IADE algorithm not only can significantly improve the convergence speed and convergence precision, but also has very good robustness, so that IADE algorithm can meet the requirements for the real-time, accuracy and stability of process optimization.

关 键 词：微分进化 全局优化 控制参数自适应 收敛速度 鲁棒性 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]