针对蚁群优化算法的关键步骤——信息素轨迹更新过程进行了深入分析。通过理论上的证明和实验验证,提出了信息素轨迹更新中存在着一个利用-探索困境;在此基础上针对这个现象提出了一种基于Metropolis接受准则的信息素更新策略,并通过在不同规模的TSP上的实验,证明了这种新策略的有效性。
An insight into the key procedure of ant colony optimization algorithm was provided. A phenomena called exploration-exploitation dilemma in the pheromone trail update was originally proposed on the basis of the theoretical arguments and experimental results. Hence a novel pheromone trail update strategy was presented based on the principle of Metropolis rule. Experiments on TSP instances with various dimension fully proved the validity of this strategy.