GRASP for WDM network design problem with traffic grooming

Two important problems arise in WDM network planning: network design to minimize the construction cost and traffic grooming to maximize the usage of high capacity channels. They are usually settled simultaneously and are denoted as the network design problem with traffic grooming (NDG). This paper proposes an approach based on Greedy Randomized Adaptive Search Procedure (GRASP) to solve the NDG problem. A mathematical formulation of the NDG problem is presented and a set of benchmarks generated according to real application scenarios are given for future research in this area, which are missing in the previous literature. Computational results show the effectiveness of the proposed GRASP algorithm by comparison with the public software CPLEX and the lower bound. This paper also shows an analysis on explaining the essential ingredients of the algorithm.