约束法蚁群算法在多目标VRP中的研究

Constraint Method Based Ant Colony System for Multi-Objective Vehicle Routing Problem

多目标车辆路径问题(MVRP)在物流研究领域具有重要的理论和现实意义,但由于各目标之间的相互联系和制约使得建模和求解具有很大的难度。在众多求解方法中,蚁群算法对解决类似组合优化问题具有明显的优势,蚁群算法已成功应用于一系列单目标优化问题,但对多目标问题的研究还处于起步阶段。侧重结合目标约束法与蚁群算法来研究多目标车辆路径问题,使各优化目标之间形成既彼此独立,又相互联系和制约的机制,最终求得多目标优化意义下的一种平衡解。仿真结果证明该算法具有良好的收敛性和运行效率,对于物流运输的实际运作具有重要的现实意义。

Multi -objective vehicle routing problem （MVRP） is very important and practical in logistic research fields, but it＇ s very difficult to model and solve because objectives have complicated relationship and restriction. Ant Colony algorithm has more obvious advantage to solve such kind of combinatorial optimization problems than many other algorithms, it has been applied successfully for solving a series of single objective optimization problems but seldom in multi - objective＇ s. This paper aims to research the multi - objective vehicle routing problem with integration of Constrain method and Ant Colony System （ACS） , form a frame in which the objectives are mutually independent and contacting each other, and find a balanceable solution for multi - objective problem finally. Experimental results show that the approach has a good astringency and high searching efficiency, and is of great importance to the logistics delivery.