文章分析了一种存在三种属性节点的物流网络,即在该网络中存在一类节点同时承担供应、需求和中转的功能。同时,还考虑了在任意两节点间单位运费可变条件下,如何进行中转点选址和流量分配,使物流网络的运输费用最小。在此背景下建立了选址分配优化模型,设计出一种将串形编码和矩阵编码相结合的两层遗传算法对其进行求解,并通过一个实例对算法加以验证并对求解结果进行分析。研究结果表明:相比于一般的需求点就近分配原则,本文提出的基于产品流进行的物流网络选址与分配模型,能够避免某些运输线路上的过度集货的现象,从而进一步降低整个物流网络的运输费用。
In this paper, we investigate a kind of logistics network with two operational characteristics. One isthat there exists a so-called triple-attribute node, i. e. , one node simultaneously performs the supply, demandand intermediate roles. The other is that the unit shipment cost between any two nodes varies over shipment vol-ume. Under this condition, to minimize the total shipment cost, we formulate a location and allocation model.Then, a two-stage Genetic Algorithms is developed to solve the problem, which combines the string coding andmatrix coding. Finally, we present a case study to explain how to solve this kind of problem with the proposedalgorithm. The results show that comparing to the general solution (i. e. , assigning the demand node to its nearbysupply or intermediate node), the proposed solution (i. e. , locating and allocating the entire logistics networkfrom the production flow's view) can avoid the excessive consolidation problems in some transportation routes.As such, it can reduce the total shipment cost for the logistics network.