自适应加速差分进化算法  

Adaptive Accelerating Differential Evolution

作　　者：许小健[1] 黄小平[1] 钱德玲[1] 

机构地区：[1]合肥工业大学土木建筑工程学院,合肥230009

出　　处：《复杂系统与复杂性科学》2008年第1期87-92,共6页Complex Systems and Complexity Science

摘　　要：差分进化算法是一种新的进化计算技术,具有良好的优化性能,但是对于高维多模态函数,算法易早熟收敛;其优化性能受差分进化模式类型及演化控制参数取值的影响较大。为此,提出自适应加速差分进化算法,该算法利用混沌的遍历性产生初始群体,以克服种群体初始化时的盲目性和随机性;其次随着搜索过程的进行随机自适应地调整缩放因子和选取差分进化模式,以减少人为因素影响,增强搜索能力。通过对多个函数进行仿真试验研究,结果表明该方法寻优效果显著,明显减少了迭代次数,提高了计算效率。Differential evolution （DE） is a new evolutionary computation technology and exhibits good performance on optimization. However the algorithm, to the high dimension and high multi-modal function, will fall into premature convergence. And its performance is strongly influenced by the differential strategy and the value of each strategy parameter including scale factor. Therefore, Adaptive Accelerating Differential Evolution （AADE） which is proposed to solve the optimization problems. The basic principle of AADE is that chaos initialization is adopted to improve individual quality, the scale factor and differential strategy adjusted randomly generation by generation to avoid the artificial factors, to enhance the searching capacity. Experiments on benchmark functions show that AADE outperform standard DE in function optimization.

关 键 词：差分进化算法 混沌初始化 差分进化模式 缩放因子 函数优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] N94[自动化与计算机技术—控制科学与工程]