基于退火的蚁群算法在连续空间优化中的应用

研究了蚁群算法在连续空间的函数寻优问题。通过修改蚂蚁信息素的留存方式和行走规则,定义了一个连续空间的蚁群算法。模拟蚂蚁用触角交流信息的过程提出了直接通信的学习机制,增强了蚂蚁的搜索能力。为了防止出现"早熟"现象,在局部搜索过程中嵌入了模拟退火的思想。同时为避免过大的残留信息,选择了新的信息增量计算函数。实例运算证明了算法的有效性。

Application of ant colony algorithm based on simulated annealing to continuous space optimization

An ant colony algorithm applied to continuous problems is proposed.This algorithm is defined by modifying both the “trail remaining” and the transfer rules.Based on the processes that ants exchange information through antennas，a novel study strategy“direct communication” is presented，which enhances the ants’ ability to search the continuous space.In the meantime，a strategy of simulated annealing is embedded in the algorithm to improve the optimization performance and prevent “premature” phenomena during the local searching.In order to avoid the large residual information，the new information increment function is applied.Experimental results show that the proposed algorithm is effective.