Development and Application of Hybrid (NLP-GA) in the Design and Operation of Pumping Stations



Pumps are among the early instruments invented by humans to be taken advantage of in the use of water utilities. Design and operation problem of pumping stations is a type of mixed integer non linear programming, in which gradient based optimization methods are not efficient for solving complex problems of this type. In this regard, such evolutionary optimization algorithms as genetic algorithm (GA) have been employed in many scientific fields of application and specially design-operation of pumping stations as a search and optimization tool. Apart from all the capabilities and advantags of such algorithms, they need a large amount of execution time as well as converging to the near optimal solutions, which are considered as some of these methods' disadvantages. This paper addresses the optimal design and operation of a pumping station system using a hybrid method for optimization consists of nonlinear programming (NLP) method and GA (NLP-GA). After evaluating the developed hybrid model in a test example mathematical problem, it has been applied and examined in a real world design-operation optimization problem. Then, the results of the NLP-GA algorithm are compared with those of three other methods for the same problem, such as GA, honey-bee mating optimization (HBMO) algorithm, NLP and the Lagrange multipliers (LM) method. The results show that the hybrid NLP-GA performers better results and faster convergence in terms of number of function evaluations toward the optimal solution than those of other methods.