Genetic algorithms applied to the design of large powerdistribution systems

摘要

This paper presents the application of a new genetic algorithm fornthe optimal design of large distribution systems, solving the optimalnsizing and locating problems of feeders and substations using thencorresponding fixed costs as well as the true nonlinear variable costs.nIt can be also applied to single stage or multistage distributionndesigns. The genetic algorithm has been tested with real sizendistribution systems achieving optimal designs in reasonable CPU timesncompared with respect to the dimensions of such distribution systems. Onnthe other hand, these distribution systems present significantly largernsizes than the ones frequently found in the technical literature aboutnthe optimal distribution planning. Furthermore, original operators ofnthe genetic algorithm have been developed in order to obtain globalnoptimal solutions, or very close ones to them. An integer codificationnof the genetic algorithm has also been used to include several relevantndesign aspects in the distribution network optimization