带佳点交叉算子的非均匀窗口蚁群算法

Ant Colony Algorithm with Good-Point Crossover Operator Based on Different Size Window

基本蚁群算法具有较强的鲁棒性,但收敛慢并容易陷入局部最优。针对这些缺陷,通过将蚂蚁的搜索空间缩减在非均匀的小窗口中,减少了蚂蚁的搜索时间。并将佳点集遗传算子引入到解的优化中来,提出了带佳点杂交算子的非均匀窗口蚁群算法,从本质上探索蚁群算法的寻优能力。实验结果表明:新提出的算法明显快于基本蚁群算法,佳点集杂交算子对解的优化有较好的作用。但需要继续探索避免陷入局部最优的方法,以及算法各部分所采用的方法的平衡问题。

Basic ant colony algorithm has strong robusmess, but has slow convergence and easily be trapped in a local optimum. Aiming at these disadvantnges, by restricting the searching space of ants in a different size small window, has a big decrease of the searching time. By a good-point set genetic operator is introduced into the optimizing of solution, proposes an ant colony algorithm with good- point crossover operator based on different size window, exploring the ability of searching best solution of ACA in essential. Experiment shows that new algorithm is obviously fast than basic ant algorithm, and good point crossover operator is benefit to optimization of solution. But it need to further explore the metthad of avoiding trapping into local optimum, and the balance of method, which is used in every part of algorithm.