Greenhouse Climate Control Simulation Data for Reinforcement Learning

Supporting information for a reinforcement learning framework optimizing greenhouse climate control. The data includes simulated metrics for crop parameters like stem elongation and resource consumption costs for heating and CO2. It was authored by Iman Hindi and last updated in March 2026.

Use Cases

• Benchmark reinforcement learning algorithms like TD3 and PPO using the provided reward function based on stem elongation and resource costs.
• Train predictive models for internal greenhouse climate using the simulated climate and control features identified by the feature-selection mechanism.
• Analyze the trade-off between crop yield growth rates and resource consumption metrics like irrigation and electricity costs.
• Simulate greenhouse environments to test control strategies for regulating temperature and CO2 concentration setpoints.

Strengths

• Data supports a novel RL framework evaluated with multiple algorithms including TD3, PPO, DDPG, and SAC.
• Includes key simulated crop parameters such as stem thickness and cumulative trusses.
• Framework was retrained under stochastic weather conditions to enhance robustness.

Limitations

• Data is simulated, not collected from physical greenhouse operations, limiting real-world validation.
• Unknown row count and column structure within the 68.3 MB ZIP file.
• Scope is limited to the specific greenhouse control simulation described in the paper.

Provenance

Source

figshare

Collection Method

Simulated data generated for a reinforcement learning research framework.

Time Range

null

Freshness

Last updated March 2026.

Geography

null