求解约束优化问题的改进教-学优化算法  

Improved Teaching-learning Based Optimization Algorithm for Constrained Optimization Problems

作　　者：李会荣[1] 周亚妮[2] 

机构地区：[1]商洛学院数学与计算机应用学院,陕西商洛726000 [2]商洛职业技术学院公共基础部,陕西商洛726000

出　　处：《海南大学学报（自然科学版）》2015年第4期333-339,共7页Natural Science Journal of Hainan University

基　　金：陕西省自然科学基础研究计划项目(2014JM2-6098);陕西省教育厅科研计划(15JK1221);商洛学院博士团队服务地方科技创新与经济社会发展能力提升专项(SK2014-01-22)

摘　　要：提出了一种非线性约束优化问题改进的教-学优化算法,该算法首先提出了自适应的教学因子,对学习阶段的迭代方程进行改进,引入了差分变异策略;其次利用约束违反度函数将约束优化问题转化为无约束双目标优化问题,在每次迭代中按照约束违反度的大小保留部分性能较优不可行个体,有效地维持了种群的多样性;最后数值实验表明,该算法具有较快的收敛速度和较好的全局寻优能力.In our report,an improved Teaching-Learning-Based Optimization（ TLBO） algorithm for constrained optimization problems was proposed. Firstly,the adaptive teaching factor was proposed,which modified the iterative equation of learner phase and introduced the mutation strategy in the differential evolution algorithm; Secondly,the constraint violation function was used to convert the constrained optimization problems into unconstrained bi-objective optimization problem,in each iteration,keeping a part of the performance of better infeasible individuals is to maintain the diversity of the swarm. The numerical experiments showed that the proposed algorithm has faster convergence speed and better ability of global optimization.

关 键 词：教-学优化算法 约束优化 差分变异 教学因子 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]