基于高斯变异的蚁狮算法及其在组合优化中的应用

Gaussian mutation based ALO and its application in combinatorial optimization

针对蚁狮(ant lion optimizer,ALO)算法在寻优后期种群数量减少、精英蚁狮影响权重减小导致算法收敛速度较慢且易陷入局优的问题,提出基于高斯变异的蚁狮(Gaussian mutation based ALO,GALO)算法。首先引用Kent混沌对初始蚂蚁种群进行扰动,提高蚂蚁种群多样性作为蚁狮寻优的基础;其次在精英蚁狮的位置更新方式中引入上一代精英蚁狮,提高算法全局搜索的能力,并通过动态切换概率平衡算法局部和全局探索的能力;最后引入高斯变异的方法,加强后期算法跳出局部最优的能力。通过10个测试函数来评估算法的寻优能力,并将其应用到0-1背包问题、桁架尺寸和动力学优化问题中,验证了GALO算法应用于组合优化问题中收敛速度更快、精度更高,为结构优化提供了一种新的方法。

Targeting to the slow convergence and potential local optimum result of the basic ant lion optimizer(ALO)algorithm,caused by the degradation of population size and elite ant lions’influence weight during the later optimizing stage,a Gaussian mutation based ALO(GALO)algorithm was proposed.First,Kent chaos was used to perturb the initial ant population and increase the diversity of the ant population,which laid the foundation of ant lion optimization.Second,the previous generation of elite ant lions was introduced into the position update method of elite ant lions to improve the algorithm’s global searching ability.Dynamic switching probability was adopted to balances the algorithm’s local and global exploration capabilities;finally,Gaussian mutation method was introduced to strengthen the ability of jumping out of the local optimum during the later optimizing stage.Ten test functions were implemented to evaluate the optimization ability of GALO.The algorithm was also applied to 0-1 knapsack problem,truss size and dynamics optimization problems,verifying that GALO achieved faster convergence and higher accuracy when applied to combinatorial optimization problems and provided a new method for structural optimization.