Offline reinforcement learning for greenhouse temperature control through natural ventilation
DOI:
https://doi.org/10.17979/ja-cea.2026.47.13686Keywords:
Data-driven control, Intelligent control system, Enviromental systems controlAbstract
Greenhouse climate control is a complex problem due to its nonlinear dynamic and the influence of strong external disturbances. This work proposes a control strategy based on Reinforcement Learning (RL) trained in offline mode to regulate the temperature of a Mediterranean greenhouse through natural ventilation. To this end, the Deep Deterministic Policy Gradient (DDPG) algorithm is employed, trained using experience collected from a real system under the operation of a proportional-integral controller (PI) considered as an expert system. The resulting policies are evaluated using data not used for learning. The results show that the agent accurately replicates the behavior of the expert controller, generating smooth control signals consistent with the system dynamics, with a control effort reduction of 7.4 % compared to the PI. This work is limited to the offline training and validation of the agent using real process data, without considering its direct deployment on the plant at this stage.
References
Castro, A. J., López-Rodríguez, M. D., Giagnocavo, C., Gimenez, M., Céspedes, L., Calle, A. L., Gallardo, M., Pumares, P., Cabello, J., Rodríguez, E., Uclés, D., Parra, S., Casas, J., Rodríguez, F., Fernandez-Prados, J. S., Alba- Patiño, D., Expósito-Granados, M., Murillo-López, B., Vasquez, L. M., Valera, D. L., 2019. Six collective challenges for sustainability of almer´ıa greenhouse horticulture. International Journal of Environmental Research and Public Health 16 (21).
García-Mañas, F., Hägglund, T., Guzmán, J. L., Rodríguez, F., Berenguel, M., 2024. A practical solution for multivariable control of temperature and humidity in greenhouses. European Journal of Control 77, 100967.
Gil, J. D., Del Rio Chanona, E. A., Guzmán, J. L., Berenguel, M., 2026. Reinforcement learning meets bioprocess control through behavior cloning: Real-world deployment in an industrial photobioreactor. Engineering Applications of Artificial Intelligence 164, 113326.
González, R. A., Ramos-Teodoro, J., Rodríguez, F., Berenguel, M., 2025. Multi-resource scheduling of the water–energy–food nexus in agroindustrial environments. In: Ocampo-Martinez, C., Quijano, N. (Eds.), Energy Systems Integration for Multi-Energy Systems: From Operation to Planning in the Green Energy Context. Springer, pp. 179–204.
Li, K., Shi, J., Hu, C., Xue,W., 2025. The intelligentization process of agricultural greenhouse: A review of control strategies and modeling techniques. Agriculture 15 (20).
Lillicrap, T. P., Hunt, J. J., Pritzel, A., Heess, N., Erez, T., Tassa, Y., Silver, D., Wierstra, D., 2015. Continuous control with deep reinforcement learning. arXiv preprint arXiv:1509.02971.
Montoya-Ríos, A. P., García-Mañas, F., Guzmán, J. L., Rodríguez, F., 2020. Simple tuning rules for feedforward compensators applied to greenhouse daytime temperature control using natural ventilation. Agronomy 10 (9).
Rajasekhar, N., Radhakrishnan, T., Samsudeen, N., 2025. Exploring reinforcement learning in process control: a comprehensive survey. International Journal of Systems Science, 1–30.
Rodríguez, F., Berenguel, M., García-Mañas, F., Guzmán, J. L., Sánchez-Molina, J. A., 2024. A multilayer control architecture for greenhouse crop production in agro-industrial districts: Conceptual framework, prospects and challenges. Smart Agricultural Technology 9, 100657.
Rodríguez, F., Berenguel, M., Guzmán, J. L., Ramírez-Arias, A., 2015. Modeling and Control of Greenhouse Crop Growth. Springer.
Sachio, S., del Rio-Chanona, E. A., Petsagkourakis, P., 2021. Simultaneous process design and control optimization using reinforcement learning. IFAC-PapersOnLine 54 (3), 510–515.
Svensen, J. L., Cheng, X., Boersma, S., Sun, C., 2024. Chance-constrained stochastic mpc of greenhouse production systems with parametric uncertainty. Computers and Electronics in Agriculture 217, 108578.
Sánchez-Molina, J. A., Gil, J. D., González, R. A., García-Mañas, F., Muñoz, M., 2025. Agroconnect research station: General description and control challenges. In: Maestre, J. M., Ocampo-Martinez, C. (Eds.), Control Systems Benchmarks. Springer, pp. 29–43.
Yusuf, A. G., Al-Yahya, F. A., Saleh, A. A., Abdel-Ghany, A. M., 2025. Optimizing greenhouse microclimate for plant pathology: challenges and cooling solutions for pathogen control in arid regions. Frontiers in Plant Science 16.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Daniel Pérez-Sánchez, Juan D. Gil, Francisco García-Mañas, Francisco Rodríguez, Manuel Berenguel
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
