A fast modeling and optimization scheme for greenhouse environmental system using proper orthogonal decomposition and multi-objective genetic algorithm

摘要

As a semi-closed ecosystem, environmental factors of greenhouses are mutual coupling, spacial distributed, and with high uncertainty. Considering the optimal environment for crop growth with energy efficiency, the optimization schemes of the greenhouse systems are studied in this paper. Different from current optimization methods, most of which are based on expert experience and parameter learning, we introduce Proper Orthogonal Decomposition (POD) technique for environmental parameters' description. On this basis, a fast optimization scheme for greenhouse environmental system is proposed. In this method, several low-dimensional parameter subspaces of greenhouse environment are constructed using POD technique. They may be embedded into optimization loop for fast solving environment response. In our case study, NSGA-II algorithm is applied for the optimization of a real greenhouse's environment. The multiple objectives include crop area's temperature distribution, carbon dioxide concentration and related energy consumption. Results show that the proposed optimization strategy has low computation cost and high space resolution, which can effectively improve the crop growth's environmental performances and save the energy consumption as well.