Pipeline Network Optimization using Hybrid Algorithm between Simulated Annealing and Genetic Algorithms

Abstract

The pipeline network is one of the most complex optimization problems consisting of several elements: reservoirs, pipes, valves, etc. The pipeline network is designed to deliver water to consumers by considering the demand and adequate pressure on the water pipe network. The main problem in designing reliable pipelines is the cost. The amount of cost that most influences the design of pipelines is the diameter of the pipe used. Therefore, this study aims to combine (hybrid) simulated annealing algorithm with genetic algorithm to optimize water pipe networks. The simulated annealing algorithm is the main algorithm in finding the optimal cost.Meanwhile, the genetic algorithm will assist in the pipeline update process using the roulette wheel selection. Simulation data is used to test the hybrid algorithm performance compared to the standard simulated annealing algorithm. The results show that the simulated annealing hybrid algorithm is able to get a more optimal cost in designing a water pipe network compared to the standard simulated annealing algorithm. Keywords: Optimization, Epanet 2.0, Simulated Annealing, and Genetic Algorithm