基于蚁群算法的离散救援问题出救点选址研究

为解决应急物流中的出救点选址问题，建立了相应数学模型，引入蚁群算法解决问题。多数应急物流可以归为点对点的支援问题，出救点的设置应该在保证出救有效的条件下使出救点最少、救援时间最短，属于双层规划问题。双层规划问题是NP难题，可以应用蚁群算法解决。出救点选址问题在蚁群算法中可以视为蚁群的聚类，通过对信息素衰减及相邻蚂蚁的吸引作为启发因子，可以得到蚁群的聚类效果。实验结果表明，基于蚁群算法的选址问题解决方案能获得理想的选址效果，收敛速度较快。

Research on depot location of discrete emergency aid based on ACO

Improved mathematic model and imported ant colony algorithm to solve the emergency location problem. Most problem of emergency logistics could be summarized as supporting from peer-to-peer. The setting of retrieval depots (RDS) should guarantee the validity of rescue, minize the number of depots, shorten the time to rescue, which was a bilevel programming problem(BPP). A BPP is theoretic NP-hard, can be solved by ant colony optimization (ACO). In ACO the RDS problem can be treated as clustering of ants. It is easy to get cluster of ants in ACO by pheromone weakening and using ants inter-attraction as elicitation. Experiments show that solution of RDS problem by ACO can get perfect effect with high speed.