最大-最小蚂蚁系统( MMAS)只在最优解对应的路径上更新信息素，有效地利用了最优解，但容易导致搜索过早停滞。文中分析了MMAS在求解车辆路径问题( VRP)时的表现，针对其容易陷入局部最优解、全局搜索能力差、后期收敛速度慢等不足提出改进，给出一种新的信息素更新策略，动态改变挥发系数的数值，并在较优的几条路线上进行信息素更新，从而在加速算法收敛的同时提高全局搜索能力，避免过早停滞。 VRP仿真实验结果表明，改进后的算法稳定性好，收敛速度比原始MMAS算法有明显的提高。%To exploit the best solutions found during an iteration or during the run of the algorithm,Max-Min Ant System ( MMAS) al-lows the ant on the best solution to heighten the pheromone. Unfortunately,it will lead to the premature stagnation of the search. By ana-lyzing the performance of MMAS in Vehicle Routing Problem (VRP),in order to avoid getting a local optimum solution,poor global search optimization ability,and slow convergence rate,a new strategy for pheromone updating is presented. It changes the value of the volatilization coefficients dynamically and updates the pheromones on the best ways,thus accelerating convergence and avoiding prema-ture stagnation. The simulation experiments of the VRP show that the stability and convergence rate of the proposed algorithm is improved significantly compared with the basic MMAS.
展开▼
