基于拉格朗日松弛算法分布式供应链优化.pdfVIP

200 6年3 月 控 制 工 程 Mar . 2 0 0 6 第13卷第2期 Control Engineering of China Vol. 13, No. 2 :2006) 周 威, 金以慧 ( , 100084) : 为解决分布环境下的无协调中心的供应链 产计划的协调问题, 提出了一种基 于拉格朗日松弛算法的折扣价格协调优化策略针对企业计划只能基于本地信息的特点, 利 用拉格朗日松弛算法将企业之间的物料耦合约束松弛掉, 从而把整 个供应链计划问题分解为 多 个可利用本地信息求解的企业 产计划子问题通过上下游企业之间对折扣价格(拉格朗日 算子) 的异步更新, 可以逐步获取整 个供应链 产计划的优化解, 从而实现分布环境下的供应 链 产计划的异步协调仿真实验证明了该方案的可行性 : 供应链; 协调; 产计划; 拉格朗日松弛; 代理次梯度 : TP 14 : A Coordination Method for Distributed Supply Chain Planning Based on Lagrangian Relaxation ZHOU Wei, JIN Yihui ( Department of Automation, Tsinghua University, Beijing 100084, China) Abstract: To the supply chain planning problem without a coordination center, a decentralized asynchronous coordination method based on La grangian relaxation algorithm is presented. By relaxing the material flow balance constraints among the enterprises, the whole supply chain plan ning problem is decomposed into multiple singleenterprise planning subproblems, which can be solved with the local information. So, each enterprise production planning model can be set up and solved independently with the discountprices, ie. Lagrangian multipliers, which are given by the upstream or downstream enterprise. To obtain the feasible solution, a distributed heuristics algorithm is proposed. During the coor dination process, through iteratively updating the discountprices among the enterprises, the near optimal solution can be achieved. The compu tational experiments show that the method can solve the supply chain planning proble