Heuristic Determination of Distribution Trees

Abstract

Electrical distribution networks develop configurations that deviate from the original long-term plan. The Distribution Trees Problem (DTP) is one means of measuring this development, which finds the deviation between long-term planning and the optimal topology for the actual conditions of the network. Each feasible solution corresponds to a set of directed out-trees rooted at the substations. DTP takes into account characteristics of the substations and consumer demand. It also determines the optimal topology of the network to distribute electrical energy at minimum cost. In this paper, we use two search techniques to solve this problem: 1) simulated annealing and 2) tabu search. Nine different problems within 500 to 30 000 consumer points and 20 substations were used to calibrate the parameters of both methods and to compare their efficiency. The numerical results indicate that the efficiency of simulated annealing decreases as the problem size increases, and that tabu search is more efficient than simulated annealing.