6th International ICST Conference on Communications and Networking in China

Research Article

Cost-effective Multilayer Network Optimization:A top down decomposition solution

  • @INPROCEEDINGS{10.1109/ChinaCom.2011.6158304,
        author={Hongfang Yu and Xiaoning Zhang and Li Wang and Vishal Anand},
        title={Cost-effective Multilayer Network Optimization:A top down decomposition solution},
        proceedings={6th International ICST Conference on Communications and Networking in China},
        publisher={IEEE},
        proceedings_a={CHINACOM},
        year={2012},
        month={3},
        keywords={multilayer optimization decomposition lagrange relaxation},
        doi={10.1109/ChinaCom.2011.6158304}
    }
    
  • Hongfang Yu
    Xiaoning Zhang
    Li Wang
    Vishal Anand
    Year: 2012
    Cost-effective Multilayer Network Optimization:A top down decomposition solution
    CHINACOM
    IEEE
    DOI: 10.1109/ChinaCom.2011.6158304
Hongfang Yu,*, Xiaoning Zhang1, Li Wang1, Vishal Anand2
  • 1: UESTC
  • 2: The College at Brockport, State University of New York
*Contact email: yuhf2004@gmail.com

Abstract

Nowadays MPLS is widely adopted as a highly scalable, protocol agnostic, data-carrying mechanism from both mobile backhauls and the optical core network. The joint optimization of IP/MPLS layer and optical layer can reduce network operating expenditures. However, the computational overhead involved in such a joint optimization is not always feasible. In this paper, we study the multilayer network optimization problem of minimizing the cost, and propose an efficient decomposition method based on Lagrange relaxation. Our work differs from previous works in that our proposed approach uses a decomposition method that decomposes the original two-layer mathematical optimization problem into an IP/MPLS-layer and an optical-layer optimization problem by relaxing the constraints between the two layers. This decomposition method enables to control the trade-off between running time and quality of the feasible solution. Numerical results for a variety of networks indicate that our proposed decomposition method can find efficient and quick solutions close to the near optimal solutions.