融合海马策略的教与学优化算法

Teaching-learning-based Optimization Algorithm with Sea-horse Optimizer

教与学优化算法(TLBO)是模拟班级内教师教学和学生学习过程的一种新型智能优化算法。为了克服教与学优化算法局部能力不强和进化后期容易陷入局部最优的问题,提出一种融合海马策略的教与学优化算法(SHOTLBO)。该算法在“教学阶段”引入海马捕食行为,并通过自适应参数的非线性变化调节SHOTLBO算法的搜索区域,增强教师的教学能力,提高种群最优解的精准性;在“学习阶段”添加学生主动学习算子,学生主动向教师学习,引入海马的两种运动行为,即levy飞行分布函数和布朗运动,实现算法大跨度随机移动,跳出局部最优解,维持种群的多样性,提升算法的全局搜索能力。在23个标准测试函数上的数值实验表明,与海马优化算法(SHO)、黑翅鸢优化算法(BKA)、TLBO、MTLBO、CMDEATLBO及RLTLBO算法相比,提出的SHOTLBO算法不仅具有较好的全局搜索能力,而且在收敛速度、收敛精度上均有明显提升。

Teaching-learning-based optimization(TLBO) is a new type of intelligent optimization algorithm that simulates the process of teachers' teaching and students' learning in the class.In order to overcome the problems of weak local ability of TLBO and easy to fall into local optimal in the late stage of evolution,a new teaching-learning-based optimization with sea-horse optimizer(SHOTLBO) is proposed.The algorithm introduces hippocampal predation behavior in the "teaching phase" and adjusts the search area of the SHOTLBO through the nonlinear change of adaptive parameters to enhance the teaching ability of teachers and improves the accuracy of the optimal solution of the population.In the "learning stage",students' active learning operators are added to actively learn from teachers,and two kinds of movement behaviors of hippocampus,namely levy flight distribution function and Brownian motion,are introduced to realize large-span random movement of the algorithm,jump out of the local optimal solution,maintain the diversity of the population,and improve the global search ability of the algorithm.Numerical experiments on 23 standard benchmark functions show that compared with SHO,BKA,TLBO,MTLBO,CMDEATLBO and RLTLBO,the proposed SHOTLBO not only has better global search ability,but also has obvious improvement in convergence speed and convergence accuracy.