particle swarm optimization for hydraulic analysis of water distribution systems

the analysis of flow in water-distribution networks with several pumps by the content model may be turned into a non-convex optimization uncertain problem with multiple solutions. newton-based methods such as gga are not able to capture a global optimum in these situations. on the other hand, evolutionary methods designed to use the population of individuals may find a global solution even for such an uncertain problem. in the present paper, the content model is minimized using the particle-swarm optimization (pso) technique. this is a population-based iterative evolutionary algorithm, applied for non-linear and non-convex optimization problems. the penalty-function method is used to convert the constrained problem into an unconstrained one. both the pso and gga algorithms are applied to analyse two sample examples. it is revealed that while gga demonstrates better performance in convex problems, pso is more successful in non-convex networks. by increasing the penalty-function coefficient the accuracy of the solution may be improved considerably.