Service restoration in distribution systems is usually formulated as a multi-objective and multi-constrained optimization problem. In order to improve the performance of Evolutionary Algorithms (EAs) applied in such problem, a new tree encoding, called Node-depth Encoding (NDE), has been successfully applied together with both a conventional EA and a modified version of the Non-Dominated Sorting Genetic Algorithm-II (NSGA-II). The objective of this paper is to verify what is the best choice to use to treat service restoration problem in large scale distribution systems? NDE with the conventional EA or NDE with the modified version of the NSGA-II. In order to do that, simulation results in two distribution systems are presented. One of these systems is the fairly large distribution system of Sao Carlos city in Brazil with 3,860 buses, 532 sectors, 509 normally closed sectionalizing switches, 123 normally open tie-switches, 3 substations, and 23 feeders.