Research on key technologies of multi-dimensional resource cooperative allocation for a super base station system
Manli Qian · The Sydney eScholarship Repository (The University of Sydney) · 2013
To meet the ever increasing demand for high wireless transmission data rates, more and more base stations (BSs) need to be deployed in current cell-centric cellular networks to increase the system capacity. This brings big challenges to mobile operators such as increased power consumption and low BS resource utilization. To solve these problems, a new centralized cellular network architecture, the super BS, is proposed by the Institute of Computing Technology, Chinese Academy of Sciences. In this architecture, large scale centralized BS resource pools are constructed, where multi-dimensional radio resource as well as the computing resource can be dynamically shared and managed across the whole network. Thus one of the key issues is how to efficiently allocate radio resource and computing resource in the super BS system. This thesis focuses on cooperative multi-dimensional resource allocation in the super BS system. A global centralized radio resource allocation algorithm based on adaptive soft frequency reuse (ASFR) is proposed, so that the system throughput is maximized. As the system traffic load fluctuates dramatically in both time and spatial domains, to avoid the high system computing overhead caused by frequent global resource allocation adjustment, a traffic aware radio resource allocation algorithm (TA-SFR) with regional optimization is then proposed. Simulation results show that both the ASFR and TA-SFR achieve a higher system throughput than traditional fixed allocation schemes. To improve system computing resource utilization, dynamic computing resource allocation algorithms are further proposed, based on an improved first fit decreasing method (IFFD), and the heuristic and simulated annealing method (HSA). Simulation results show that compared to the fixed allocation in traditional cellular networks, the IFFD and HSA effectively decrease system power consumption by 40%-65% with high traffic load and 70%-100% with low traffic load.