Optimization of irrigation scheduling using ant colony algorithms and an advanced cropping system model

Abstract

A generic simulation-optimization framework for optimal irrigation and fertilizer scheduling is developed, where the problem is represented in the form of decision-tree graphs, ant colony optimization (ACO) is used as the optimization engine and a process-based crop growth model is applied to evaluate the objective function. Dynamic decision variable option (DDVO) adjustment is used in the framework to reduce the search space size during the generation of trial solutions. The framework is applied for corn production under various levels of water availability and rates of fertilizer application in eastern Colorado, USA. The results indicate that ACO-DDVO is able to identify irrigation and fertilizer schedules that result in better net returns while using less irrigation water and fertilizer than those obtained using the Microsoft Excel spreadsheet-based Colorado Irrigation Scheduler (CIS) tool for annual crops. Another advantage of ACO-DDVO compared to CIS is the identification of both optimal irrigation and fertilizer schedules.