求解约束优化问题的协同进化教与学优化算法  

A Co-evolutionary Teaching-learning-based Optimization Algorithm for Constrained Optimization Problems

作　　者：刘三阳 靳安钊 LIU San-Yang;JIN An-Zhao(Department of Mathematic Science,Xidian University,Xi＇an 710126)

机构地区：[1]西安电子科技大学数学与统计学院,西安710126

出　　处：《自动化学报》2018年第9期1690-1697,共8页Acta Automatica Sinica

基　　金：国家自然科学基金(61373174)资助~~

摘　　要：对约束优化问题,为了避免罚因子和等式约束转化为不等式约束时引入的约束容忍度参数所带来的不便,本文在基本教与学优化(Teaching-learning-based optimization,TLBO)算法中加入了自我学习过程并提出了一种求解约束优化问题的协同进化教与学优化算法,使得罚因子和约束容忍度随种群的进化动态调整.对7个常见测试函数的数值实验验证了算法求解带有等式和不等式约束优化问题的有效性.In order to avoid the inconvenience of penalty factors and the tolerance amount during transforming equality constraints into inequality constraints, the self-learning process is combined with teaching-learning-based algorithm, and a co-evolutionary teaching-learning-based algorithm is thus proposed, which makes the penalty factors and tolerance amounts dynamically adjust along with the population evolution. Numerical experiments on seven common test functions verify the effectiveness of the algorithm to solve optimization problems with equality and inequality constraints.

关 键 词：教与学优化算法 协同进化 罚因子 约束容忍度 种群多样性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]