Enhancing the Efficiency of the Automatic Design of Rural Drainage Networks

Abstract

The discharges flowing through rural drainage networks depend not only on the local climatic and hydrologic characteristics, but also on the geometric characteristics of the channels constituting the network. For this reason, the evaluation of the design discharges should be accomplished during the search of the optimal network. This leads to time-consuming optimization procedures, and it is desirable to devise efficient numerical alternatives. Two novel models, EGA and EGA-f, are proposed in order to increase the numerical efficiency of genetic algorithms (GAs) for the solution of the optimal rural drainage network problem. Both EGA and EGA-f procedures are based on the use of the nodal excavation depths at the channel ends as decision variables. Moreover, the EGA-f procedure improves EGA by freezing temporarily the design discharges during the optimization process in the case where uniform flows through the channels can be assumed. The application of the two models is demonstrated by means of numerical experiments, confirming their superiority over existing GAs for the optimization of rural drainage networks.