Global performance of metaheuristic optimization tools for water distribution networks

Numerous metaheuristic optimization algorithms are used for optimal design of water distribution networks. Each algorithm shows dissimilar characteristics depending on the network properties and the sensitivity analysis of the algorithm control variables. New performance metrics of metaheuristic optimization methods are proposed using simple but robust refined metrics and were applied to the available literature data for different algorithms which have previously been used for three popular benchmark water distribution networks. In general, recent performance metrics are devoted to measure effectiveness, efficiency, and reliability in a separate manner, which made some confusion, which is the best?. In the present work, the proposed metrics are used to calculate both of best global and average global performance of different optimization algorithms. The results show that the present metrics have a good distinctive performance between different algorithms. The Fittest individual referenced Differential Evolution is found to be the best algorithm.