基于增广Lagrangian算法的供应链生产计划协调优化  

Coordination method for multi-plant supply chain planning optimization based on augmented Lagrangian relaxation

作　　者：周威[1] 金以慧[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《清华大学学报（自然科学版）》2005年第10期1324-1327,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(60174046)

摘　　要：为解决具有分布特点的多厂供应链生产计划的协调问题,提出了一种基于增广L agrang ian松弛算法的内部价格协调优化策略。利用增广L agrang ian松弛算法将工厂之间的耦合约束松弛,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单厂计划子问题。为获取问题可行解与加快算法的收敛速度,又分别提出了一种前溯式可行化算法与一种模糊次梯度算法。通过协调中心对产品内部价格的迭代更新,实现了整个供应链生产计划的协调优化。仿真结果表明,该策略能够较好地协调多厂供应链计划,效果明显优于已有的协调方法。Traditional centralized optimization methods can not easily solve the multi-plant supply chain planning problem with a distributed decision-making structure. A decentralized coordination method based on the augmented Lagrangian relaxation algorithm was developed by relaxing the material flow balance constraints between plants so that the supply chain planning problem is decomposed into multiple single-plant planning sub-problems, which can be solved with the local information. Each plant production planning model can then be solved independently with the inner-prices, i.e. the Lagrangian multipliers, given by the coordination center. The solution is obtained using a distributed forward heuristics algorithm. The near optimal solution can be achieved during the coordination process by iteratively updating the inner-prices in the coordination center. Computational tests show that the method can efficiently solve the multi-plant supply chain planning problem.

关 键 词：供应链 协调 生产计划 增广Lagrangian松弛算法 

分 类 号：TP14[自动化与计算机技术—控制理论与控制工程]