引入侦查子群的二进制蚁群算法求解函数优化问题

Binary Ant Colony Algorithm with Scouting Subgroup for Function Optimization Problem

为求解函数优化问题,将遗传算法中的二进制编码方式引入标准蚁群算法.但由于该算法迭代过程中易出现早熟停滞现象,为此提出一种改进的蚁群算法,该算法在原算法基础上引入一定比例的侦查蚁群.侦查蚁群以一定的概率做侦查搜索以扩大解的搜索空间;在信息素更新策略上,为兼顾当代和历代的搜索成果,采取信息素混合更新策略,同时增强侦查子群的最佳路径信息及其余蚁群的路径信息.最后,通过对几个经典测试函数的求解,证明该算法解决函数优化问题非常有效,不仅能够克服早熟现象,而且能够加快收敛速度.

To solve function optimization problem,this work introduces binary code of genetic algorithm to standard ant colony algorithm.But this algorithm often gets stuck into premature stagnation during the iteration process.Therefore,this paper proposes an improved algorithm with a scouting subgroup at a certain ratio.It can expand the searching space for solution by random search-route of scouting subgroup.And hybrid pheromone＇s updating strategy is adopted to make use of both current-fruit and precedent-fruit.The pheromone of the best-scouting subgroup＇s route as well as others＇ is increased.At last,the improved algorithm was tested for a series of classical functions.The results show that it can not only handle these optimization problems effectively,but also can avoid pre-maturity and accelerate constringency.