An Analysis of the Solution Representation in an ILS Algorithm Applied to a Network Equipment Location Problem

This work analyzes the influence of the solution representation on the performance of an algorithm based on Iterated Local Search applied to a problem of location and allocation of network equipment. A new solution representation is proposed, accompanied by the necessary adaptations in the procedures for building the initial solution, evaluation function, local search and disturbance mechanism, maintaining the general structure of the algorithm to allow a direct comparison with the previous approach. The two representations were evaluated in different instances, considering metrics such as the best solution found, processing time, and variability between executions. The results indicate that both representations achieve high-quality solutions; However, the new solution representation presents significant gains in computational efficiency, greater robustness and less variability of the results. These results reinforce the importance of choosing the representation of the solution in the design of meta-heuristic algorithms.