多策略集成的哈里斯鹰算法求解全局优化问题

Multi-strategy Integrated Harris Hawk Algorithm toSolve Global Optimization Problems

为提高哈里斯鹰算法优化问题的性能,提出一种融合佳点集、非线性能量逃逸因子和Logistic-Cubic级联混沌扰动的多策略哈里斯鹰优化算法(Improve Harris Hawk Optimization,IHHO)。首先,引入佳点集策略代替随机初始种群,均匀初始种群分布性。其次,根据算法各个阶段不同特征提出一种非线性能量逃逸因子,平衡全局和局部勘探能力。最后,引入Logistic-Cubic级联混沌对搜索位置扰动,避免算法陷入局部最优。利用IHHO算法求解23个函数及三桁架工程设计问题,并利用目标收敛曲线、Wilcoxon秩和检验进行测试,结果表明,IHHO算法相比对比算法具有更强寻优性能、求解稳定性,在求解全局优化问题上具有一定竞争性。

In basic science and practical engineering applications,there are problems of solving optimization schemes in different dimensions and under multiple constraints,and it is difficult for most conventional methods to deal with such optimization problems effectively.Intelligent optimization algorithms are able to solve many optimization problems in the case of failure of classical optimization techniques due to their small search space,few search times,flexible computation and strong applicability.With the widespread application of intelligent optimization algorithms in many fields such as logistics scheduling,combinatorial optimization,system control,etc.,more and more scholars have begun to study such algorithms.Most of these algorithms are designed by the influence of biological and physical phenomena in nature,and are increasingly used in the engineering field due to their advantages of simple concept and easy implementation.Harris Hawk Optimization(HHO)was proposed in 2019 as a new type of intelligent optimization algorithm,which was inspired by the hunting behavior of the Harris Hawk,and has the advantages of simple principle,easy programming,fewer parameters,high convergence accuracy and fast convergence,and has been applied to the design and engineering optimization problems in certain disciplines.For different types of function optimization problems and engineering applications,the HHO algorithm has the problems of slow convergence speed and insufficient stability of optimization search.In order to further improve the performance of the Harris Hawk algorithm in solving problems,this paper proposes a multi-strategy Improved Harris Hawk Optimization(IHHO)algorithm that integrates the good point set,nonlinear energy escape factor,and Logistic-Cubic cascading chaotic perturbations.Firstly,for the characteristics of random generation of the initial population,the good point set strategy is applied for optimization to uniformly distribute the initial population and improve its traversal ability.Secondly,for the problem that the algorithm is easy to fall into local optimal solution,a nonlinear energy escape factor is proposed based on the different characteristics of each stage of the algorithm,and the escape factor changes from large to small according to the number of iterations,i.e.,expanding the search range in the early iteration to prevent the algorithm from falling into local optimal,and reducing the search range in the late iteration to accelerate the convergence of the algorithm,so as to balance the algorithm’s global and local exploration ability.Finally,for the problem that the search position is easy to converge locally,Logistic-Cubic cascade chaos is introduced to perturb the search position during the updating process,to avoid the algorithm from falling into local optimum,and to improve the solution accuracy and convergence speed.In the simulation experiment stage,the IHHO algorithm is used to solve 23 function problems with different characteristics,each problem is solved 30 times,and the mean and standard deviation of each result is taken to compare with the other 7 algorithms.And the results are verified by using the target convergence curve and Wilcoxon rank sum test.The results indicate that the IHHO algorithm has stronger optimization performance and solution stability than other algorithms.At the same time,the IHHO algorithm is used to optimize the solution of the three-truss design engineering problem,and the results show that the algorithm has strong competitiveness compared to the comparative algorithms,and has the ability to become an effective tool for solving global optimization problems.In the future,further research will be conducted on intelligent optimization algorithms,combining them with deep reinforcement learning to further solve larger and more complex practical application problems.