Thesis or Dissertation Optimization Models for Robust and Power Efficient Networks


pp.1 - 78 , 2016-06-24 , The University of Electro-Communications
With the ever increasing traffic demands in the networks, it has become essential to design networks that are robust to different traffic conditions. The common approach of proceeding by over-provisioning to accommodate traffic change, or events such as link failures, leads to a massive waste of resource and energy. On the other hand, it is challenging to provide solutions which are robust to traffic change due to unpredictable nature of the traffic. This thesis focuses on providing robust optimization models for power efficient and resilient networks with traffic uncertainty. In the first part of the thesis an optimization model for power efficient networks with traffic uncertainty is introduced. Most of the studies nowadays pertaining to the field of green communications are based on estimates of real traffic matrix. However predicting the traffic matrix is a difficult task for network operators. These models may not be fully applicable in a context where the traffic often fluctuates. The optimization approach presented in the thesis, which is named the hose model with bound of link traffic (HLT), is extended from the hose model. By using HLT, the knowledge of the exact traffic information is not required. The model is constructed by specifying the total outgoing and incoming amount at each node and the total traffic going through each link. The power efficiency problem is formulated as a mixed integer linear programming (MILP) problem, with an objective to reduce the flow through each link and allow the links to be put to sleep mode. To mitigate the complexity of solving the MILP formulation, an heuristic based on a linear programming formulation of HLT is presented. Simulation results show that the scheme, while being robust to traffic uncertainty, achieves power efficiency comparable to the models where the knowledge of the traffic information is required. In the second part of the thesis, the optimization model based on HLT for resilient networks, is presented. The spare capacity assignment problem is investigated. Preplanned link restoration preallocates protection resources during the configuration phase of the networks. The technique involves over-provisioning of the network resources to provide enough spare capacity in order to route the traffic on backup paths in case of failure in the primary network. As a consequence, this induces extra-costs for the operators. Minimizing the toll incurred by the protection techniques is desirable for them. Most of the link protection techniques nowadays are either designed to provide along the backup path(s) the same capacity as the primary failed link, or by taking into account exact values of the traffic demands. While it is possible to reconsider the capacity requirement, it is worth recalling the difficulty for the operators to know the exact traffic matrix (set of traffic demands) in the network. We provide a mathematical formulation to minimize the capacity requirements of the backup networks, which considers the traffic uncertainty. The design of a backup network, which exclusively reroutes the traffic in case of a link failure in an existing primary network, is considered. We formulate a linear programming problem for the design and capacity provisioning of backup networks with traffic uncertainty under HLT. From the simulation results, a reduction ranging from 29.75% to 70% is achieved in the spare capacity required for the backup networks.

Number of accesses :  

Other information