Abstract
AbstractScheduling of the rehabilitation activities of water main networks depends mainly on available budget and planning time. Other factors such as network reliability, criticality, location, contract size, and rehabilitation method(s) affect the optimization of the scheduling process. This paper presents a method for optimized scheduling of rehabilitation work for water distribution networks. The method utilizes unsupervised neural networks (UNNs) and mixed-integer nonlinear programming (MINLP) and performs the scheduling in two stages. In the first stage, UNNs are used to cluster water mains into groups according to locations and rehabilitation methods of water mains. In the second stage, MINLP is used to determine the number of rehabilitation contract packages and the generation of optimized scheduling of these packages considering network reliability, criticality, contract size, and planning time. In order to demonstrate the essential features of the developed method, a case study was analyzed and the results obtained are discussed; highlighting their utilization in practice.