Soft-computing methods in greenhouse environmental and crop modelling

This paper presents results regarding ongoing experimental research in a hydroponic greenhouse located at the University of Algarve in the South of Portugal. The work focuses on environmental predictive control and production optimisation for a tomato crop. Recent results on the application of Soft- Computing methods to external climate predictive models, and also on strategies for tomato growth modelling are presented. The climate and environmental quantities are modelled by using radial basis function neural networks whose input-output structure is selected by a multi-objective genetic algorithm. Image processing methods are applied to the extraction of plant features from pictures taken periodically