ADVANTAGES OF USING GENETIC ALGORITHMS OVER DETERMINISTIC METHODS IN OPTIMAL DESIGN OF THE WATER NETWORKS

Abstract. Water supply and distribution systems are key elements of any urban development or sustainable agricultural development in the case of the irrigation systems. It is not surprising then that various analyses and design techniques for the economic and efficient design of hydraulic networks have received attention over the years. A possible approach to increasing the hydraulic capacity of water supply and distribution systems is to add loops (thereby providing alternative pathways) and increase their hydraulic capacity with minimum capital investment. However, optimal design of looped hydraulic networks belongs to the class of large combinatorial optimisation problems which are very difficult to handle using conventional operations research techniques. In recent years some successful attempts to apply modern heuristic methods to this problem have been published. The main part of the paper deals with applying such a technique, namely the genetic algorithm methodology, to network rehabilitation optimisation considering both technical and economic aspects of the problem. A case study of the sprinkler irrigation system Šala-Kolárovo 13 (south-west part of Slovakia) is presented. 1. Introduction. The linear programming (LP) method has long been accepted as an approach for the optimal selection of diameters for pipes in branche