A Parallel Multiobjective Artificial Bee Colony Algorithm for Dealing with the Traffic Grooming Problem

This work presents a novel parallel multiobjective approach based on the Artificial Bee Colony algorithm for grooming low-speed traffic requests onto high-capacity optical channels. The traffic grooming problem in mesh optical networks is an NP-hard problem, so the usage of metaheuristics and parallelism jointly for increasing the network performance is a great option in order to reduce execution times. The parallel multiobjective approach is implemented by using OpenMP. We have measured the speedup and efficiency obtained by our parallel approach with 2, 4, 8, and 16 cores. Efficient numerical results are reported in the experimental phase conducted on two optical networks. Finally, we present a comparative study with traditional methods; in which we show that the usage of swarm intelligence outperforms previous results published in the literature.