Abstract: To reasonably design the distribution network of enterprises, the location-distribution problem in two-stage distribution network was proposed. In a single product's two-stage distribution network which contained plant, distribution center and customer, the location of distribution center with known numbers and the delivery plan from plant to distribution center and from distribution center to customer were determined in many candidate sites. On the premise of satisfying capacity limit of plant and distribution center, customer demand and distribution center's minimum number of transshipment and distribution, the total cost which included location fixed cost of distribution center and delivery cost from plant to distribution center and from distribution center to customer was minimized.A mixed integer linear programming model was constructed, and a combined 0-1 and priority-based encoding genetic algorithm was designed. Computational tests on a case study and a set of randomly generated instances with 2 to10 plants, 5 to 100 distribution centers, 20 to 400 customers and 6 to 20 distribution centers to be located were conducted to demonstrate the feasibility and effectiveness of proposed method.

Key words: two-stage distribution network, location-distribution, mixed integer linear programming model, genetic algorithms, boundary constraints

摘要： 为合理设计企业分销网络，提出两级分销网络选址—配送问题，研究在由工厂、配送中心和客户组成的且只销售一种产品的两级分销网络中，从若干候选点里确定已知个数的配送中心的位置以及工厂到配送中心和配送中心到客户的配送方案，使得在满足工厂和配送中心能力限制、客户产品需求、配送中心最小转运量和最小配送量的前提下，包括配送中心选址固定费用、从工厂到配送中心和配送中心到客户的配送费用在内的总费用最小，构建该问题的混合整数线性规划模型，并设计0-1和优先权混合编码的遗传算法。通过对1个演示算例和1组包括2～10个工厂、5～100个配送中心候选点、20～400个客户、6～20个需建设配送中心数的随机生成算例的计算分析，表明了所提方法的可行性和有效性。

关键词: 两级分销网络, 选址—配送, 混合整数线性规划模型