Performance and Cost Trade-off in Tracking Area Reconfiguration: A Pareto-optimization Approach

Article English OPEN
Modarres Razavi, Sara ; Yuan, Di ; Gunnarsson, Fredrik ; Moe, Johan (2012)
  • Publisher: Ericsson Research, Ericsson AB, Sweden
  • Related identifiers: doi: 10.1016/j.comnet.2011.08.017
  • Subject: Bi-criteria optimization; Reconfiguration; Signaling overhead; Tracking Area | Telecommunications | Telekommunikation

Tracking <img src="" />Area<img src="" /> (TA) design is one of the key tasks in location management of Long Term Evolution (LTE) networks. TA enables to trace and page User Equipments (UEs). As UEs distribution and mobility patterns change over time, TA design may have to undergo revisions. For revising the TA design, the cells to be reconfigured typically have to be temporary torn down. Consequently, this will result in service interruption and “<img src="" />cost<img src="" />”. There is always <img src="" />a<img src="" /> trade-<img src="" />off<img src="" /> between the <img src="" />performance<img src="" /> in terms of the overall signaling overhead of the network and the <img src="" />reconfiguration<img src="" /><img src="" />cost<img src="" />. In this paper, we model this trade-<img src="" />off<img src="" /> as <img src="" />a<img src="" /> bi-objective <img src="" />optimization<img src="" /> problem to which the solutions are characterized by <img src="" />Pareto<img src="" />-optimality. Solving the problem delivers <img src="" />a<img src="" /> host of potential trade-offs among which the selection can be based on the preferences of <img src="" />a<img src="" /> decision maker. An integer programming model has been developed and applied to the problem. Solving the integer programming model for various <img src="" />cost<img src="" /> budget levels leads to an exact scheme for <img src="" />Pareto<img src="" />-<img src="" />optimization<img src="" />. In order to deliver <img src="" />Pareto<img src="" />-optimal solutions for large networks in one single run, <img src="" />a<img src="" /> Genetic Algorithm (GA) embedded with Local Search (LS) is applied. Unlike many commonly adopted <img src="" />approaches<img src="" /> in multi-objective <img src="" />optimization<img src="" />, our algorithm does not consider any weighted combination of the objectives. Comprehensive numerical results are presented in this study, using large-scale realistic or real-life network scenarios. The experiments demonstrate the effectiveness of the proposed <img src="" />approach<img src="" />. funding agencies|CENIIT||Swedish Research Council||Linkoping University||
Share - Bookmark