Abstract: This paper investigates a novel bi-level inventory and distribution planning model with defective item disposal over multiple periods under a fuzzy random environment. In the bi-level model, the advanced logistics provider is the leader in the hierarchal process that decides purchase quantities to minimize total global cost. In the lower level, the inventory and distribution service provider aim to minimize costs in his branch and make decisions based on the advanced logistics provider’s decisions. Considering the complexity of the model, a Bi-level Dynamic Programming-based Particle Swarm Optimization with a multiple social learning structures（Bi-DPGLNPSO） algorithm with fuzzy random simulation is presented to solve the model. A case is given to demonstrate the efficiency and effectiveness of the proposed model and proposed Bi-DPGLNPSO algorithm.

Key words: inventory distribution management, defective item disposal, Particle Swarm Optimization（PSO）, bi-level programming, fuzzy random

摘要： 针对考虑残次品的多生产商选择多商品多阶段库存配送问题，建立了一个基于动态规划的双层库存配送模型。高端物流服务集成商以整个供应链网络成本最小为目标制定采购决策；库存配送服务商以运营成本最小为目标，在集成商决策下制定库存和配送决策。设计了模糊随机环境下基于动态规划的双层全局-局部-邻域粒子群算法（Bi-DPGLNPSO）对模型进行求解。并通过算例验证模型和算法的有效性和合理性。通过参数测试和算法对比检验算法的优越性。

关键词: 库存配送管理, 残次品处理, 粒子群算法, 双层规划, 模糊随机