ADVANCED MODEL-BASED GREENHOUSE CLIMATE CONTROL USING MULTI-OBJECTIVE OPTIMIZATION

Modern greenhouse climate control requires use of advanced climate-control models; however, adoption of advanced climate-control models in today’s industrial greenhouse production is hindered by the shortcoming of existing climate-control systems to support non-invasive composition of independently-developed climate-control models. Despite the fact that adoption of advanced climate-control models allows growers to optimize their production through improved energy efficiency, improved plant quality and yield as well as reduced risks for various climate-related diseases, commercial vendors of industrial greenhouse-climate-control systems have not taken action to provide the necessary support for independent extensibility in their systems so far. Present climate-control systems require the control logic of independently-developed climate-control models to be merged into a single monolithic climate-control model. Hence, addition of new climate-control models requires modification and validation of this monolithic model. In this paper, we present a new approach to extensible greenhouse climate control that allows new climate-control models to be added dynamically to the climate-control system independently of each other. There is no need for merging models into a single monolithic model, as the approach allows independently-developed models to co-exist alongside each other. The novelty of the approach is the use of a genetic algorithm to compute a balanced greenhouse climate that satisfies the multi-objective-optimization problem defined by the independently-added climate-control models. Feasibility of the approach is demonstrated through simulation of a number of selected production scenarios using a generic greenhouse simulator. The results of the simulations clearly show that the approach finds a balanced greenhouse climate that is satisfactory to the requirements of the independent climate-control models.