Towards Constraint Optimal Control of Greenhouse Climate

Greenhouse climate is a multiple coupled variable, nonlinear and uncertain system. It consists of several major environmental factors, such as temperature, humidity, light intensity, and CO, concentration. In this work, we propose a constraint optimal control approach for greenhouse climate. Instead of modeling greenhouse climate, Q-learning is introduced to search for optimal control strategy through trial-and-error interaction with the dynamic environment. The coupled relations among greenhouse environmental factors are handled by coordinating the different control actions. The reinforcement signal is designed with consideration of the control action costs. To decrease systematic trial-and-error risk and reduce the computational complexity in Q.-learning algorithm Case Based Reasoning (CBR) is seamlessly incorporated into Q-learning process of the optimal control. The experimental results show this approach is practical, highly effective and efficient.