Multi-Objective Genetic Algorithm for Optimal Scheduling of Chlorine Dosing in Water Distribution Systems


This paper presents two new multi-objective genetic algorithm models using a Pareto-based selection technique for determining the optimal schedule of chlorine dosing within a water distribution system with multiple, competing objectives: primarily disinfection control and aesthetic control. Formulating the optimal dosing problem using a Pareto-based multi-objective genetic algorithm offers water utilities an innovative solution technique, with each optimisation simulation returning a Pareto-optimal set of candidate solutions. An overview of the object-oriented application architecture, encapsulating each model, is also given, including the independent linked design with the water network simulation program. Both models were applied to a hypothetical distribution system, including comparison with existing non-Pareto techniques. Results are discussed, including the advantages of using the new Pareto-based multi-objective models.