Genetic algorithm design of piped irrigation systems

Abstract

The design of piped irrigation systems is a complex matter. In its simplest form it involves the selection of the sizes of pipes and other hydraulic elements so as to ensure that the varying demands for irrigation water can be met while maintaining suitable operating pressures throughout the system. The hydraulic design of a system usually involves a trial-and-error procedure with sizes being selected based on experience and judgement and the performance of the system being assessed using a hydraulic computer simulation package (such as KYPIPES or EPANET). Genetic algorithms (GAs) are a class of stochastic optimization techniques which enable near-optimal solutions to be identified for complex engineering systems. This paper describes the application of the genetic algorithm technique to the design of piped irrigation systems. The results obtained indicate that significant savings (in the order of 16% of capital cost) can be achieved by the application of this technique