Midterm Umut Selcuk

Improving Constraint Compliance in Economically Optimal Greenhouse Climate Control under Weather Uncertainty

Supervisor: Dr. Koty McAllister

Abstract: Greenhouses are an ancient technology in agriculture that allow for considerable control over the farming environment. The climate inside modern greenhouses can be manipulated through climate control systems to directly impact the crop quality and yield. Advancements in control theory and greenhouse modelling have made receding horizon optimal control a viable strategy for economically optimal climate management. This approach balances crop yield gains against the costs of climate control. Nominal MPC controllers generate efficient trajectories that often operate near constraint boundaries, leading to violations under weather uncertainties. Accounting for uncertainties in the weather forecast to eliminate constraint violations comes with computational difficulties. This thesis investigates methods to improve constraint compliance in greenhouse MPC while keeping computational requirements reasonable, offering a pathway to more robust and cost-effective climate management.