Evaluation of the Performance and Parameter Sensitivity of Constraint-Handling Techniques for a Genetic Algorithm-Based Irrigation Canal Network Management Problem

Handling constraints in Evolutionary Algorithms (EAs) can be challenging, especially for highly constrained, complex optimization problems. Because constraints are not handled directly in an EA, heuristics must be carefully designed and applied to guide the search toward feasible, good-performing solutions. A common way to handle constraints in EAs is to use penalty functions. These penalty functions can have various formulations, and often require time-consuming fine-tuning of the penalty parameters to effectively search for feasible solutions with good objective function values. This paper investigates the performance of various constraint-handling techniques applied to an EA-based irrigation canal network management problem. The problem involves determining efficient network operations that minimize the water consumed while preventing problems such as water shortage and flooding. The case study is based upon a complex large-scale irrigation network in Egypt. The constraint-handling methods investigated include three penalty functions. The associated fine-tuning parameters for each constraint-handling method are varied to investigate the sensitivity of the solution quality (feasibility and objective function value) to these parameters. The solution quality and its sensitivity to the constraint-handling parameters are compared for the constraint-handling methods. This comparison is intended to help provide guidelines for better performing penalty function methods for an EA-based irrigation canal network management problem that can be successfully implemented without excessive fine-tuning requirements.