Abstract
In the face of increased competition for water resources, optimal irrigation scheduling is necessary for sustainable development of irrigated agriculture. However, optimal irrigation scheduling is a nonlinear problem with many competing and conflicting objectives and constraints, and deals with an environment in which conditions are uncertain. In this study, a multi-objective optimization problem for irrigation scheduling was presented in which maximization of net benefits and water use efficiency and minimization of risk were the objectives. The presented optimization problem was solved using four different approaches, all of which used the AquaCrop model and nondominated sorting genetic algorithm III. Approach 1 used dynamic climate data without adaption; Approach 2 used dynamic climate data with adaption; Approach 3 used static climate data without adaption; and Approach 4 used static climate data with adaption. The dynamic climate data were generated using the bootstrap resampling of original climate data. A case study of maize production in north Jiangsu Province of China was used to evaluate the proposed approaches. Under the multi-objective scenario presented and other conditions of the study, Approach 4 gave the best results, and showed that irrigation depths of 400, 325, and 200 mm were required to produce a maize crop in a dry, normal, and wet year, respectively.