A comparative study of deep reinforcement learning for crop 1 production management

Authors: BALDERAS, JOSEPH - University Of Texas; CHEN, DONG - Mississippi State University; Huang, Yanbo; WANG, LI - University Of Texas At Arlington; LI, REN-CHENG - University Of Texas At Arlington

Submitted to: Smart Agricultural Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/20/2025
Publication Date: 3/3/2025
Citation: Balderas, J., Chen, D., Huang, Y., Wang, L., Li, R. 2025. A comparative study of deep reinforcement learning for crop 1 production management. Smart Agricultural Technology. 10(2025):2772-3755. https://doi.org/10.1016/j.atech.2025.100853.
DOI: https://doi.org/10.1016/j.atech.2025.100853

Interpretive Summary: Artificial intelligence (AI) has been expected to enhance decision support to crop management for optimizing yield and minimizing a field’s environmental impact to crop fields. The research has been conducted to investigate and promote an innovative AI approach, reinforcement learning (RL), to learning of optimal decision-making strategies through trial and error in dynamic environments or developing adaptive crop management policies. The studies have shown that RL can generate crop management policies that compete with, and even outperform, expert-designed policies within simulation-based crop models.

Technical Abstract: Crop production management is essential for optimizing yield and minimizing a field’s environmental impact to crop fields, yet it remains challenging due to the complex and stochastic processes involved. Recently, researchers have turned to machine learning to address these complexities. Specifically, reinforcement learning (RL), a cutting-edge approach designed to learn optimal decision-making strategies through trial and error in dynamic environments, has emerged as a promising tool for developing adaptive crop management policies. RL models aim to optimize long-term rewards by continuously interacting with the environment, making them well-suited for tackling the uncertainties and variability inherent in crop management. Studies have shown that RL can generate crop management policies that compete with, and even outperform, expert-designed policies within simulation-based crop models. In the gym-DSSAT crop model environment, one of the most widely used simulators for crop management, proximal policy optimization (PPO) and deep Q-networks (DQN) have shown promising results. However, these methods have not yet been systematically evaluated under identical conditions. In this study, we evaluated PPO and DQN against static baseline policies across three different RL tasks, fertilization, irrigation, and mixed management, provided by the gym-DSSAT environment. To ensure a fair comparison, we used consistent default parameters, identical reward functions, and the same environment settings. Our results indicate that PPO outperforms DQN in fertilization and irrigation tasks, while DQN excels in the mixed management task. This comparative analysis provides critical insights into the strengths and limitations of each approach, advancing the development of more effective RL-based crop management strategies.