Optimal Processing and Trading for a Commodity in the Presence of Inventory Conversion Flexibility and Random Supply

• 成果类型: Article
• 署名作者: Bhandari, Ashish S.; Sapra, Amar; Seshadri, Sridhar; Sastry, Trilochan
• 署名单位: Indian Institute of Management (IIM System); Indian Institute of Management Bangalore; University of Illinois System; University of Illinois Urbana-Champaign; University of Illinois System; University of Illinois Urbana-Champaign
• 刊物名称: PRODUCTION AND OPERATIONS MANAGEMENT
• ISSN/ISSBN: 1059-1478
• DOI: 10.1177/10591478251331127
• 发表日期: 2025
• 页码: 3138-3156
• 关键词: Commodity Trading; Cooperative; Random Supply; dynamic programming; Blending

摘要

This paper presents a decision support system used by an agricultural cooperative in the Indian states of Andhra Pradesh and Telangana to optimize the purchase, blending, sale, and storage of groundnuts for maximum profit. The cooperative buys raw groundnuts (input commodity) from member farmers and processes them into multiple grades of groundnut seeds (output commodity). These may then be blended to create intermediate grades to exploit arbitrage opportunities. The cooperative sells part of the output on the spot market while storing the rest for future periods. A key challenge is the random supply of input commodity-driven by the cooperative's obligation to accept all member produce-and the option to blend the output. Unlike prior work, this study examines blending across a multiperiod planning horizon, a novel aspect in operations management literature. The problem is modeled as a dynamic program over a harvest season. We analyze the structure of the optimal value function and decisions and find that the function is not separable in input and output inventories, which complicates the identification of optimal solution. However, in special cases such as when blending is disallowed, the function simplifies. An efficient computational procedure is developed for the general case. Using real cooperative data, we demonstrate that multiperiod blending significantly boosts profits-by 100-900%, or Indian National Rupees 1.94-17.46 million annually-highlighting the value of this approach.