Abstract: To overcome the premature convergence that frequently appears in the extended Ant Colony Optimization (ACO), a new hybrid method was presented, which mutated individuals by reverse mutation and differential evolution after every step in the evolution of extended ant colony algorithm. Evolutionary process of the extended ant colony algorithm would be affected by the calculation information of the reverse mutation and differential evolution algorithm. As a result, the diversity of population was maintained and the global optimization would be realized. The numerical results indicate that the proposed algorithm has high precision and strong robustness.

Key words: Ant Colony Optimization (ACO), extended ant colony optimization, reverse mutation, Differential Evolution (DE)

摘要： 为了克服扩展蚁群优化(ACO)容易出现早熟现象，提出在扩展蚁群每次进化后引入倒序变异和差分进化对新种群进行二次变异。通过倒序变异和差分进化(DE)算法计算的信息来影响扩展蚁群的进化进程，以保持群体的活性，实现全局优化的目的。数值试验结果表明新算法精度较高、鲁棒性较强。

关键词: 蚁群优化, 扩展蚁群优化, 倒序变异, 差分进化