Genetic Algorithms in Irrigation Planning: A Case Study of Sri Ram Sagar Project, India

Abstract

The present study deals with application of Genetic Algorithms (GA) in the field of irrigation planning. The GA technique is used to achieve efficient operating policy with the objective of maximum net benefits for the case study of Sri Ram Sagar Project, Andhra Pradesh, India. Constraints include continuity equation, land and water requirements, crop diversification considerations, and restrictions on storage capacities. Penalty function approach is used to convert constrained problem into unconstrained one. For fixing GA parameters , namely, crossover and mutation probabilities, the model is run for 7 values of crossover and 6 values of mutation probabilities. It is found that appropriate parameters such as number of generations, population size, crossover probability, and mutation probability are 200, 50, 0.6 and 0.01 respectively for the present study. Maximum benefits obtained by LP solution is 2.4893 Billion Rupees where as these are 2.3903 Billion Rupees by GA (with a fitness function value of 2.3678 Billion Rupees). Results obtained by GA are compared with Linear Programming solution and found to be reasonably close.