Decision support for irrigation project planning using a genetic algorithm

This work presents a model based on on-farm irrigation scheduling and the s imple genetic algorithm optimization (GA) method for decision support in ir rigation project planning. The proposed model is applied to an irrigation p roject located in Delta, Utah of 394.6 ha in area, for optimizing economic profits, simulating the water demand, crop yields, and estimating the relat ed crop area percentages with specified water supply and planted area const raints. The user-interface model generates daily weather data based on long -term monthly average and standard deviation data. The generated daily weat her data are then applied to simulate the daily crop water demand and relat ive crop yield for seven crops within two command areas. Information on rel ative crop yield and water demand allows the genetic algorithm to optimize the objective function for maximizing the projected benefits. Optimal plann ing for the 394.6 ha irrigation project can be summarized as follows: (1) p rojected profit equals US$ 114,000, (2) projected water demand equals 3.03 x 10(6) M-3, (3) area percentages of crops within UCA#2 command area are 70 .1, 19, and 10.9% for alfalfa, barley, and corn, respectively, and (4) area percentages of crops within UCA#4 command area are 41.5, 38.9, 14.4, and 5 .2% for alfalfa, barley, corn, and wheat, respectively. Simulation results also demonstrate that the most appropriate parameters of GA for this study are as follows: (1) number of generations equals 800, (2) population sizes equal 50, (3) probability of crossover equals 0.6, and (4) probability of m utation equals 0.02. (C) 2000 Published by Elsevier Science B.V.