混合改进蚁群算法的函数优化  

Function optimization based on an improved hybrid ACO

作　　者：陈明杰[1] 黄佰川[1] 张旻[1] 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《智能系统学报》2012年第4期370-376,共7页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金资助项目(51079033);中央高校基本科研业务费专项资金资助项目(HEUCF100430)

摘　　要：针对蚁群算法进化速度慢、容易出现停滞现象的不足,探讨了一种基于自适应信息素挥发因子的改进蚁群算法.针对蚁群算法容易陷入局部最优的缺点,提出了一种基于决策变量高斯变异的改进蚁群算法.针对蚁群算法速度慢的不足,探讨了一种基于决策变量边界自调整的改进蚁群算法.将上述3种改进相融合,提出了一种基于自适应信息素挥发因子、决策变量高斯变异和决策变量边界自调整3种改进策略的混合改进蚁群算法.将其应用于函数优化中,仿真结果表明,混合改进蚁群算法在收敛速度和收敛率方面都有很大改进,具有更好的寻优性能.Considering the low evolution speed and the tendency towards stagnation of ant colony optimization（ACO）,an improved ACO was discussed based on an adaptive pheromone evaporation factor.To avoid the defect of ACO easily falling into the local optimum,another improved ACO was proposed based on Gaussian variation of decision variables.To overcome the shortcoming of the slow speed of ACO,a new and improved ACO was given based on boundary self-tuning of the decision variables.Finally,an improved hybrid ant colony algorithm was proposed,which combined the adaptive pheromone evaporation factor,Gaussian variation of decision variables,and boundary self-tuning of the decision variables.When applied to function optimization,the simulation results show that the improved hybrid ACO has a higher degree of accuracy,a higher convergence ratio,and improved optimization performance.

关 键 词：混合改进蚁群算法 函数优化 自适应 高斯变异 蚁群算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]