GreenLight-Gym: Reinforcement learning benchmark environment for control of greenhouse production systems

2024-10-06

AI Summary

The absence of efficient and robust reinforcement learning (RL) benchmark environments hinders algorithm development for greenhouse crop production. Method: We introduce the first open-source RL simulation platform tailored for agricultural intelligent control. It features a differentiable C++ core implementing greenhouse dynamics, accelerated numerical integration and automatic differentiation via CasADi, and a modular Python interface supporting multi-task configuration, parametric uncertainty modeling, and standardized evaluation. Contributions/Results: (1) Simulation speed is 17× faster than pure Python implementations; (2) It enables, for the first time, end-to-end training of robust climate controllers (PPO/SAC) under parameter perturbations; (3) It provides a unified evaluation framework, filling a critical gap in standardized RL benchmarks for agriculture. The platform significantly lowers the barrier to algorithm validation and advances research and deployment of intelligent greenhouse control systems.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Intelligent Robots: Learning & Optimization for ROB
• Machine Learning: Reinforcement Learning

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Agentic search

Abstract

This study presents GreenLight-Gym, a new, fast, open-source benchmark environment for developing reinforcement learning (RL) methods in greenhouse crop production control. Built on the state-of-the-art GreenLight model, it features a differentiable C++ implementation leveraging the CasADi framework for efficient numerical integration. GreenLight-Gym improves simulation speed by a factor of 17 over the original GreenLight implementation. A modular Python environment wrapper enables flexible configuration of control tasks and RL-based controllers. This flexibility is demonstrated by learning controllers under parametric uncertainty using two well-known RL algorithms. GreenLight-Gym provides a standardized benchmark for advancing RL methodologies and evaluating greenhouse control solutions under diverse conditions. The greenhouse control community is encouraged to use and extend this benchmark to accelerate innovation in greenhouse crop production.

Problem

• Develop RL methods for greenhouse crop control
• Improve simulation speed for greenhouse models
• Standardize benchmark for diverse control conditions

Innovation

• Differentiable C++ implementation with CasADi framework
• Modular Python wrapper for flexible control tasks
• 17x faster simulation than original GreenLight model