基于配对机制的人类学习优化算法研究

摘要

Swarm intelligence algorithm is a search mechanism that simulates biological evolution or animal-group collaboration, aiming to search the solution space of complex problems quickly and efficiently, and find the global optimal solution. In recent years, by imitating the human learning mechanism, a novel swarm intelligence algorithm—the human learning optimization (HLO) algorithm, is proposed. Based on the HLO algorithm and combined with the inspiration of human mating phenomenon, an optimization algorithm for human learning based on pairing mechanism (PHLO) is proposed for the first time. The PHLO consists of four operators: random learning operator, individual learning operator, paired learning operator, and social learning operator. The 0-1 knapsack is taken as a benchmark problem, and the optimization results of the new algorithm are compared with the HLO and the simulated annealing algorithm (SA). The experimental results indicate that the PHLO is obviously superior to the HLO algorithm and SA algorithm, no matter the speed of convergence or the precision of optimization.%群体智能算法是模拟生物进化或动物群体协作的搜索机制,其目标是快速有效地搜索复杂问题的解空间,寻求全局最优解.近几年,模仿人类学习机制,提出了一种新的群体智能算法,即人类学习优化(HLO)算法.基于HLO算法,结合人类社会婚配现象获得的启发,首次提出一种基于配对机制的人类学习优化算法(PHLO).人类学习优化算法(PHLO)包含四个运算符——随机学习运算符、个体学习运算符、配对学习运算符和社会学习运算符.以0-1背包问题作为测试基准,将新算法的优化结果与HLO、模拟退火算法(SA)进行比较.实验结果表明,无论收敛速度还是寻优精度,PHLO都明显优于HLO算法和SA算法.