提出了一种具有寿命估算的最大-最小蚂蚁系统,该算法对进化种群数进行宏观调控的同时,用个体寿命限制个体的生存期,利用蚂蚁能估算自己所剩寿命的能力,实现其突变性的大风险搜索行为,从而扩大搜索范围,即主动性增加蚁群多样性,使蚁群逃离局部最优,并为实现蚁群的快速收敛提供可能。通过对TSP问题的仿真实验结果表明,该方法能够有效防止早熟收敛,较好地平衡局部搜索和全局搜索的收敛性能。
This paper proposes a variant of ACO called Max-Min Ant System with Lifetime Estimation(MMAS_LE).MMAS_LE uses both macroscopical control and microcosmic control based on lifetime estimating of ants to realize mutation-al and risky search behavior,which can broaden the search space,break local convergence and speed the global convergence by increasing the population diversity actively.The simulation results on TSP problem show that MMAS_LE can efficiently avoid premature convergence and well balance the exploration and exploitation.