A novel hybrid optimization model to determine optimum water resources for water supply of residential areas

In this study, a new hybrid model was improved by combining heuristic and numerical optimization algorithms to decide on optimum water resources based on their costs in the water supply. The purpose of the hybrid model is to reach a best result in the shortest time by simultaneously searching global and local minimums. Therefore, the steepest descent (SD) algorithm (numerical optimization method) was embedded in the classical modified clonal selection algorithm (the classical modified Clonalg) (one of artificial immune systems, heuristic optimization technique). This hybridization allows the SD algorithm to search local minimums while the classical modified Clonalg is searching a global minimum. The hybrid optimization model was applied to the cost objective function depending on distances and piezometric head differences between the water resources and destination. A scenario consists of five hypothetical water resources and one residential area/settlement. Herein, the aim is to satisfy the water demand of the residential area with a minimum cost from the water resources. The cost objective function was also minimized by the regular model (the classical model) according to the scenario, and their results were compared. Both models were run ten times for testing their stabilities. According to the results, the hybrid model is better than the regular model in terms of run time and stability. The hybrid model found a minimum cost for the water supply in a shorter time (in half) in comparison with the regular model in all runs.