Abstract:
In the present study an evolutionary based optimization algorithm, Differential Evolution (DE) and a heuristic optimization algorithm, Simulated Annealing (SA) is compared and applied to a case study of Bisalpur project, Rajasthan, India. Crops grown in the command area are Jowar, Maize, Groundnut, Soyabean, Wheat, Barley, Gram, and Mustard. Bisalpur project covers two districts, namely, Tonk and Sawai Madhopur for irrigation and Ajmer, Jaipur and Tonk districts for drinking water. The objective of irrigation planning model is to determine suitable cropping pattern, reservoir operating and release policy which yields maximum net benefits from irrigation while meeting drinking water requirements. The planning model is subjected to constraints such as continuity equation, land area, irrigation diversion, reservoir, canal capacity and crop area restrictions. Fifty percent dependable inflow level is considered for the planning scenario. Population size (NP), Crossover constant (CR), Weighting factor
(F) and ten different strategies (variations) of DE and initial temperature (T), number of iterations (n) performed at a particular temperature and cooling rate (CoR) of SA are analyzed in the present planning problem. Results of SA are compared with solution of DE. It is observed that the net benefits (crores of rupees) obtained from ten DE strategies are 95.1901,
97.722 (constraint violation), 96.027 (constraint violation), 94.997 (constraint violation),
95.1903, 95.183, 95.1903, 95.1901, 95.1211 and 95.1903 whereas it is 95.1897 in case of SA with a combination of n=150, CoR=0.95 and T=1000. Efforts were also made to analyze the CPU time for different combinations of DE and SA on a PC with PIV 2.4GHz./256MB RAM/40GB HDD configuration. It is concluded that DE/rand-to-best/I/bin is the best strategy for the planning problem with maximum net benefits of 95.1903 crores of rupees which is slightly higher as compared to SA.
