Algoritmo Híbrido combinando a Têmpera Simulada e Busca Tabu para Localização Ótima de Geradores em Redes de Distribuição Radiais

Abstract

Metaheuristics have been effective in solving large-scale combinatorial optimization problems, in which there are many control variables, constraints and candidate solutions. The optimal location of generators in distribution networks, is one of those optimization problems in which the solution space is discrete and large. In fact, location is a part of the problem called optimizing distributed generation, which also covers the sizing and dispatch of generating units.The proposal is to combine simulated annealing (SA) and tabu search (TS), two established metaheuristics, in order to better adapt them to the specific application. The power losses in the feeder are taken as an objective function to be minimized. The networks are considered radial, without this implying a generality restriction, but because it is the predominant configuration. The expected contribution of the article is in the way in which the search for the optimal solution is carried out: instead of being free throughout the solution space, it is a search in neighborhoods of the current solution. As the SA is characterized by being a global search method and the TS, a local search method, their combination showed efficiency superior to that of the separate algorithms for almost all the different distribution networks to which it was applied.