Scholarlink

This paper presents De-Supply, a novel approach utilizing deep reinforcement learning for multi-objective supply-chain optimization. It formulates the problem as a Markov Decision Process with complex action and observation spaces, addressing real-world supply chain challenges. De-Supply’s custom policy network enables agents to make informed procurement decisions, resulting in efficient stock management and cost control. Extensive experiments demonstrate its effectiveness, using historical data to outperform baselines and human-level performance. De-Supply offers robust solutions for warehouse businesses and holds promise for broader applications in the industry.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

De-Supply: Deep Reinforcement Learning for Multivariate Supply Chain Optimization

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Franck Romuald Fotso Mtope,
  • Diptangshu Pandit,
  • Sina Joneidy,
  • Farzad Rahimian

摘要

This paper presents De-Supply, a novel approach utilizing deep reinforcement learning for multi-objective supply-chain optimization. It formulates the problem as a Markov Decision Process with complex action and observation spaces, addressing real-world supply chain challenges. De-Supply’s custom policy network enables agents to make informed procurement decisions, resulting in efficient stock management and cost control. Extensive experiments demonstrate its effectiveness, using historical data to outperform baselines and human-level performance. De-Supply offers robust solutions for warehouse businesses and holds promise for broader applications in the industry.