Factorial Two-Stage Irrigation System Optimization Model

Abstract

This study proposes a factorial two-stage irrigation system optimization model (FTIM) for supporting agricultural irrigation water-resource management under uncertainty. The FTIM incorporates fractional factorial design, two-stage stochastic programming (TSP), interval linear programming (ILP), and interval probability and is applied to agricultural water allocation. The FTIM can take full advantage of conventional two-stage optimization approaches to tackle uncertainties presented as intervals, to investigate potential interactions among input parameters and their influences on system performance, and to enhance applicability to dual uncertainties expressed as interval probabilities. The proposed FTIM approach is for the first time applied to a hypothetical case study of water resource allocation in an agricultural irrigation problem. The results indicate that the effects of parameters on the objective function are evaluated quantitatively, which can help decision makers screen out significant parameters, analyze their interactions in model response, and identify possible schemes with maximized net system benefit. Especially for the study problem, the most positive significant factor affecting total net benefits is water quality at a medium flow; penalties resulting from undelivered water and benefit rates of onion farms in both periods have negative effects.