基于改进蚁群算法求解连续空间寻优问题

Ant Colony Algorithm for Solving Continuous Function Optimization Problem Based on Pheromone Distributive Function

蚁群算法是近几年优化领域中出现的一种启发式仿生类并行智能进化算法,并在离散空间领域中得到广泛应用,但在求解连续空间优化问题方面的研究相对较少。为了克服蚁群算法在连续空间中搜索时间过长等缺点,在原有的连续空间寻优方法的基础上,提出了一种用于求解连续空间寻优问题的改进蚁群算法。针对各子区间内的总信息量及应有的蚁数的求解方式进行改进,引入一个随迭代次数增加而变化的函数,以提高改进后蚁群算法的收敛速度。仿真实验表明,提出的基于信息量分布函数的改进蚁群算法较有关文献的算法有更好的收敛性能,从而为蚁群算法求解这类问题提供了一种可行有效的新方法。

Ant colony algorithm,in recent years,emerges as a novel approach of bionic meta-heuristic algorithm in the field of optimization.Though it is widely applied in the discrete space area,it is relatively less researched in solving continuous function optimization.Aiming at overcoming the shortage of long time in searching for continuous function with ant colony algorithm,the paper proposes an improved ant colony algorithm for solving continuous function optimization,which is based on the original methods of continuous function optimization.The improvement is directed against the total amount of pheromone and size of ant colony within all the subintervals.It leads-in a function that varies with increase of the iterations,in the hope of increasing the convergence speed of ant colony algorithm after its improvement.And numerical simulation results indicate that,comparing with the algorithm proposed by References,this improved algorithm offers better solution for continuous space optimization problems,hence it is an effective new way to solve problems alike.