一类用于连续域寻优的蚁群算法

由真实蚁群觅食行为启发而来的经典蚁群算法,非常适合解决组合优化问题,但经典蚁群算法的离散性本质也限制了其在连续空间问题求解中的应用。为此,提出了一种用于连续域寻优的改进蚁群算法。局部搜索上基于解决离散域问题的经典蚁群优化思想,全局搜索利用类似于遗传算法的交叉、变异操作-称为Ant Diffusion和Ant Walk方法,每代寻优结束后均采用"精英策略"把本代最优个体保留到下一代中。最后,采用改进算法对几个基准函数做了寻优测试,都取得了良好的效果,证明了算法的有效性。

An Improved Ant Colony Algorithm for Continuous Domains

To the problem that the ant colony algorithm being fit to solve the optimization problem of the discrete function,its discrete nature restricts applications to the continuous domains,an improved ant colony optimization algorithm is introduced.This is a stochastic search algorithm,which needs not continuous evaluation of derivatives for the object function.The improved ant colony approach,in the local search,is based on the idea of ACO,and the Ant Diffusion and Ant Walk method being similar to the crossover and mutation operation of genetic algorithm in the global search is utilized.And while each generation accomplishing,the best individual is preserved to next generation by using the idea of"elitist strategy".This algorithm is tested for variety of different benchmark functions,and it can handle these optimization problems very well.