Metaheuristic approach to solve a variant of generalized regenerator location problem in optical networks

Abstract

The generalized regenerator location problem (GRLP) deals with the optimal placement of regenerators in optical network in order to preserve signal quality between the end-user pairs with minimal number of installed regenerators. This study considers a variant of GRLP that involves weights of end-user pairs reflecting their importance or priorities in an optical network and the costs of installing regenerators for each location. The considered GRLP variant addresses two objectives: to maximize the sum of weights of connected end-user pairs and to minimize the total costs of installing regenerators. As the optical network involves large number of nodes, metaheuristic approach is used to solve the problem under consideration. Several S-metaheuristic and P-metaheuristic concepts are modified in accordance to the problem's characteristics. The proposed metaheuristic methods are tested on the modified GRLP data sets from the literature and the obtained results are compared in terms of solution quality in respect to several metrics.