TITLE:
A Markov Based Performance Analysis of Handover and Load Balancing in HetNets
AUTHORS:
Baoling Zhang, Weijie Qi, Jie Zhang
KEYWORDS:
Heterogeneous Networks (HetNets), Markov Chain Process (MC), Cell Range Expansion (CRE), Hysteresis Margin (HM) and Time-to-Trigger (TTT)
JOURNAL NAME:
International Journal of Communications, Network and System Sciences,
Vol.10 No.10,
October
25,
2017
ABSTRACT: LTE heterogeneous networks (HetNets) is becoming a popular topic since it was first developed in 3GPP Release 10. HetNets has the advantage to assemble various cell networks and enhance users’ Quality of Service (QoS) within the system. However, its development is still constrained by two main issues: 1) Load imbalance caused by different transmission powers for various tiers, and 2) The unbalanced transmission power may also increase unnecessary handover rate. In order to solve the first issue, Cell range expansion (CRE) can be applied in the system, which will benefit lower-tier cell during user association phase; CRE, Hysteresis Margin (HM) and Time-to-Trigger (TTT) will be utilized to bound UE within lower tier network of HetNets and therefore solve the second issue. On the other hand, the relationship of these parameters may be complicated and even reduce QoS if they are chosen incorrectly. This paper will evaluate the advantage and disadvantage of all three parameters and propose a Markov Chain Process (MCP) based method to find optimal HM, CRE and TTT values. And then, the simulation is taken and the optimal combination for our scenario is obtained to be 1 dB, 6 dB and 60 ms respectively. First contribution of this paper is to map the HetNets handover process into MCP and all the phases of handover can be calculated and analysed in probability way, so that further prediction and simulation can be realised. Second contribution is to establish a mathematical method to model the relationship of HM, CRE and TTT in HetNets, therefore the coordination of these three important parameters is achieved to obtain system optimization.