Revenue-driven Lightpaths Provisioning over Optical WDM Networks Using Bee Colony Optimization

• Published in: International journal of computational intelligence systems Vol. 10; no. 1; pp. 481 - 494
• Main Author: Marković, Goran Z.
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 2017; Springer
• Subjects: bee colony optimization (BCO); lightpath; optical network; Research Article; revenue maximization; routing and wavelength assignment (RWA); lightpath; routing and wavelength assignment (RWA); optical network; revenue maximization; bee colony optimization (BCO)
• ISSN: 1875-6891; 1875-6883; 1875-6883
• DOI: 10.2991/ijcis.2017.10.1.33

This paper aims to study the lightpaths provisioning problem in optical WDM networks with scarce available wavelengths under the static (off-line) traffic demands such that network operator’s (NO’s) revenue is maximized. To achieve this goal, a NO has to be addressed with the issue how to solve the call admission control jointly with the lightpaths routing and wavelength assignment (RWA) problem in efficient manner. The improved bee colony optimization (BCOi) metaheuristic is applied to solve the considered revenue maximization (Max-Rev) problem. We evaluated the performances of the proposed BCOi Max-Rev framework by performing numerous simulation experiments in different realistic WDM optical network topologies. We observed that our BCOi Max-Rev algorithm is an efficient tool to produce high quality solutions within reasonable amount of CPU time. It has been proved that BCOi Max-Rev solutions just slightly deviate from optimal solutions (at most 1%) and considerably outperform some heuristic algorithms, such as the Max-Profit and FCFS. In addition, our Max-Rev BCOi algorithm is able to produce better solution quality compared to the constructive BCO approach (up to 3.5% in the case of NSFNet and 5% in the case of EON). Finally, we compared the BCOi to differential evolution (DE) approach in the case of more complex networks, such as the USA optical network topology. The results show that our BCOi always outperforms DE metaheuristic, whereby the profit improvement could reach up to 20 % in some instances.