计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (4): 240-242.

• 工程与应用 • 上一篇    下一篇

基于强化学习和半马氏过程的供应链优化

杨鹏 赵辉 呼生刚   

  1. 南开大学信息学院 空军工程大学工程学院
  • 收稿日期:2006-03-07 修回日期:1900-01-01 出版日期:2007-02-01 发布日期:2007-02-01
  • 通讯作者: 杨鹏

Applying Reinforcement Learning and Semi-Markov Decision Process to Optimize Supply Chain Performance

  • Received:2006-03-07 Revised:1900-01-01 Online:2007-02-01 Published:2007-02-01

摘要: 在网络化制造环境下,供应链在地理分布上的分散性、市场需求的随机性都使得供应链的管理越来越复杂。本文应用强化学习和半马氏过程理论针对跨地区且存在地区生产成本差异的供应链管理问题进行了建模,分析了在随机需求的情况下,供应链的库存决策问题。应用实例说明本文方法的可行性和有效性。

关键词: 供应链管理, 库存控制, 强化学习, 半马氏过程

Abstract: In the networked manufacturing environment, the geographical dispersal of supply chains, and the stochastic demands of the markets increase the complexity of the system. In this paper, reinforcement learning and semi-Markov process were applied to inventory control of supply chain management ranged among regions with different production costs. The inventory decision under stochastic demands was analyzed. The simulation result showed that the proposed method is promising.

Key words: supply chain management, inventory control, reinforcement learning, semi-Markov process