Knowledge-based optimal irrigation scheduling of three-dimensional agro-hydrological systems

Abstract

An agricultural irrigation scheduler determines how much to irrigate and when to irrigate for a field. The accurate and effective scheduler decision for large agricultural fields is still an open research problem. In this work, we address the high dimensionality of the agricultural field and propose a knowledge-based approach to provide optimal irrigation amount and irrigation time for three-dimensional agro-hydrological systems. First, we introduce a structure-preserving model reduction technique to decrease the dimension of the system. Then, based on the reduced model, an optimization-based scheduler is designed. In the design of the scheduler, empirical knowledge from farmers is considered to significantly reduce the computational complexity. The proposed scheduler is designed in the framework of model predictive control. The objective of the proposed scheduler is to maximize crop yield while minimizing irrigation water consumption and the associated electricity usage. The proposed approach is applied to a field to show the effectiveness and superiority of the proposed framework.