Application of Ant-Colony-Based Algorithms to Multi-Reservoir Water Resources Problems

In this paper, the continuous Ant Colony Optimization Algorithm (ACOR) is used to investigate the optimum operation of complex multi-reservoir systems. The results are compared with those of the well-known Genetic Algorithm (GA). For this purpose, GA and ACOR are used to solve the long-term operation of a three-reservoir system in Karkheh Basin, southwestern Iran. The solution must determine monthly releases from the three reservoirs and their optimum allocations among the four agricultural demand areas. Meanwhile, a minimum discharge must be maintained within the river reaches for environmental concerns. Review of past research shows that only a few applications of Ant Colony have been generally made in water resources system problems; however, up to the time of initiating this paper, we found no other application of the ACOR in this area. Therefore, unlike GA, application of Ant-Colony-based algorithms in water resources systems has not been thoroughly evaluated and deserves serious study. In this paper, the ACOR is stuided as the most recent Ant-Colony-based algorithm and its application in a multi-reservoir system is evaluated. The results indicate that with when the number of decision variables increases, a longer computational time is required and the optimum solutions found are inferior. Therefore, the ACOR would be unable to solve complex water resources problems unless some modifications are considered. To overcome a part of these drawbacks, a number of techniques are introduced in this paper that considerably improve the quality of the method by decreasing the required computation time and by enhancing optimum solutions found.