A Real Case of Multi-Period Water Distribution Network Design solved by a Hybrid Simulated Annealing

Abstract

In this work, we propose an optimization solver based on a hybrid Simulated Annealing (HSA) to optimize the water distribution network design. The problem formulation includes multi-period restrictions with time-varying demand patterns. The HSA search process is affected by the Markov Chain Length (MCL), making modifications in the network design. For that reason, we studied the HSA behaviour by considering static and dynamic methods to compute the MCL. We test the algorithms by using networks reported in the state-of-the-art and also a real and new median size network that arises from a regional necessity. The experimentation suggests the use of a dynamic method, which exhibits the balance between efficacy and efficiency.