融合差分变异的教-学优化算法  

Teaching-learning-based optimization algorithm by using differential mutation

作　　者：李会荣[1] 乔希民[1] 赵鹏军[1] 

机构地区：[1]商洛学院数学与计算机应用学院,数学研究所,陕西商洛726000

出　　处：《计算机工程与应用》2016年第5期36-40,共5页Computer Engineering and Applications

基　　金：国家自然科学基金(No.11161001);陕西省自然科学基础研究计划项目(No.2014JM2-6098;No.2013JM023;No.2014JM1019);陕西省教育厅科研计划项目(No.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(No.SK2014-01-22);商洛学院服务地方专项(No.13SKY-FWDF001)

摘　　要：教-学优化算法(TLBO)是模拟班级中的教学过程和学习过程而提出的一种新型智能优化算法。为了改善教-学优化算法的性能,结合差分进化算法,提出一种融合差分变异的教-学优化算法(DMTLBO)。该算法提出自适应的教学因子,根据差分进化算法中变异策略,对学习阶段迭代方程进行改进,使得学习者的学习能力不仅受到学习者之间的相互影响,而且还受到当前最好学习者的影响,提高了算法的收敛速度。仿真实验表明,该算法的收敛速度和寻优精度均优于TLBO算法、PSO算法以及DE算法。Teaching-Learning-Based Optimization（TLBO）algorithm is an new intelligent optimization algorithm which simulates the teaching process and learning process of a classroom. In order to improve the performance of teachinglearning-based optimization algorithm, an improved teaching-learning-based optimization algorithm by using differential mutation（DMTLBO）is proposed. The adaptive teaching factor is proposed, the iterative equation of learner phase is modified by using mutation strategy in the differential evolution algorithm, so the learners increase their knowledge not only by the interaction between themselves, but also through input from the best learner in the current iteration. Simulation results show that DMTLBO outperforms TLBO, PSO and DE in terms of convergence speed and search precision.

关 键 词：智能优化算法 教学优化 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]