混合策略改进的金枪鱼群优化算法

Improved Tuna Swarm Optimization Algorithm Based on Hybrid Strategy

关键词:  莱维飞行|Circle混沌映射|金枪鱼群优化算法|群智能优化|基准函数

Key words:  Levy flight|Circle chaotic map|tuna swarm optimization algorithm|swarm intelligence optimization|benchmark functions

摘要:

针对金枪鱼群优化(Tuna Swarm Optimization,TSO)算法前期收敛速度慢和容易陷入局部最优等不足，提出混合策略改进的金枪鱼群优化算法(Improved Tuna Swarm Optimization Algorithm Based on Hybrid Strategy,HTSO)。首先，用Circle混沌映射初始化种群，提高种群的丰富性；其次，利用莱维飞行(Levy flight)在空间随机游走的搜索特点，提高算法在螺旋式觅食时的幅度，减少算法陷入局部最优的次数，帮助其快速找到全局最优。通过14个基准测试函数，在不同维数下比较传统TSO算法、HTSO、鲸鱼优化算法(Whale Optimization Algorithm,WOA)、灰狼优化(Grey Wolf Optimizer,GWO)算法和哈里斯鹰优化(Harris Hawks Optimization,HHO)算法的性能。仿真结果表明，不管是在低维还是在高维的情况下，HTSO比其他4种算法有更好的寻优性能和鲁棒性。最后对HTSO进行wilcoxon秩和检验，验证结果表明，HTSO与其他对比算法存在显著性差异。

Abstract:

Aiming at the shortcomings of Tuna Swarm Optimization (TSO) algorithm,such as slow convergence speed and easy to fall into local optimum,an Improved Tuna Swarm Optimization Algorithm Based on Hybrid Strategy (HTSO) was proposed.Firstly,the Circle chaotic map was used to initialize the population and improve the richness of the population.Secondly,using the search characteristics of Levy flight random walk in space,the amplitude of the algorithm in spiral foraging was improved,the number of times the algorithm falls into local optimum was reduced,and the global optimum was found quickly.Through 14 benchmark test functions,the performance of traditional TSO algorithm,HTSO,Whale Optimization Algorithm (WOA),Grey Wolf Optimizer (GWO) algorithm and Harris Hawks Optimization (HHO) algorithm was compared under different dimensions.The simulation results show that HTSO has better optimization performance and robustness than the other four algorithms in both low-dimensional and high-dimensional situations.Finally,the wilcoxon rank sum test of HTSO is carried out.The verification results show that there are significant differences between HTSO and other comparison algorithms.