Scholarlink

<p>Agricultural products are often subject to seasonal fluctuations in production and demand. Predicting and managing inventory levels in response to these variations can be challenging, leading to either excess inventory or stockouts. Additionally, the complex network of the agri-food supply chain, comprising various stakeholders, involves multifaceted interactions and dependencies. The limited shelf life of the products further adds to the complexity of the supply chain, posing challenges for implementing traditional methods to obtain an optimal set of solutions. To bridge these research gaps, this study proposes a novel Deep Reinforcement Learning (DRL) based hybrid algorithm, Asynchronous Advantage Actor-Critic with Distributed Proximal Policy Optimization (A3C-DPPO), for inventory optimization in the agri-food supply chain characterized by uncertain demand and lead times. The optimization problem is formulated as a Markov Decision Process (MDP), and by selecting optimal order quantities with a continuous action space, the proposed algorithm effectively addresses the challenges associated with dynamic decision-making. To evaluate the performance of the proposed algorithm in a real-world scenario, empirical data from the fresh agricultural products supply chain inventory is considered. The proposed algorithm addresses the challenges associated with uncertain demand and lead time, resulting in improved performance in terms of overall profit margins. The findings reveal that the A3C-DPPO algorithm significantly outperforms traditional DRL approaches by effectively handling complex state-action spaces.</p>

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

Adaptive inventory strategies using deep reinforcement learning for dynamic agri-food supply chains

  • 期刊论文

  • 发表日期：

    2026年2月11日

查看全文
  • Amandeep Kaur,
  • Gyan Prakash

摘要

Agricultural products are often subject to seasonal fluctuations in production and demand. Predicting and managing inventory levels in response to these variations can be challenging, leading to either excess inventory or stockouts. Additionally, the complex network of the agri-food supply chain, comprising various stakeholders, involves multifaceted interactions and dependencies. The limited shelf life of the products further adds to the complexity of the supply chain, posing challenges for implementing traditional methods to obtain an optimal set of solutions. To bridge these research gaps, this study proposes a novel Deep Reinforcement Learning (DRL) based hybrid algorithm, Asynchronous Advantage Actor-Critic with Distributed Proximal Policy Optimization (A3C-DPPO), for inventory optimization in the agri-food supply chain characterized by uncertain demand and lead times. The optimization problem is formulated as a Markov Decision Process (MDP), and by selecting optimal order quantities with a continuous action space, the proposed algorithm effectively addresses the challenges associated with dynamic decision-making. To evaluate the performance of the proposed algorithm in a real-world scenario, empirical data from the fresh agricultural products supply chain inventory is considered. The proposed algorithm addresses the challenges associated with uncertain demand and lead time, resulting in improved performance in terms of overall profit margins. The findings reveal that the A3C-DPPO algorithm significantly outperforms traditional DRL approaches by effectively handling complex state-action spaces.