Research On Low-complexity Resource Allocation In Heterogeneous Wireless Networks

The heterogeneous wireless networks are composed of macrocell networks and small cell networks.Resource management techniques are the main techniques to solve the problems that the heterogeneous wireless networks are facing.However,since the coverage of small cells is small and the number of small cells is large,the complexity of resource allocation for small cell networks is generally too high.To reduce the complexity,this paper mainly studies the low-complexity resource management algorithms in the heterogeneous wireless networks.My research mainly focuses on the following respects.Firstly,we concentrate on the optimization of energy efficiency in the heterogeneous wireless network.We propose a low-complexity energy-efficient power control algorithm considering interference mitigation.Firstly,we model the optimization problem as a Stackelberg game and propose a dynamic pricing mechanism to mitigate the cross-tier interference.Then by applying the exponentially-weighted low-pass filter,we obtain the closed-form of the best response of transmit power and the pricing factors,through which the complexity of the proposed algorithm can be largely decreased.Simulation results show that the proposed algorithm greatly reduces the computational complexity when compared with NPCA at the cost of slight loss of the energy efficiency.Secondly,we force on the optimization of spectrum efficiency in the heterogeneous wireless network.To provide statistical delay guarantees,EC is employed as the network performance measure instead of the conventional Shannon capacity.To reduce the complexity,we decompose the resource management problem into three sub-problems,namely,cluster configuration sub-problem,intra-cluster sub-channel allocation sub-problem and inter-cluster power control sub-problem.Firstly,we develop a low-complexity heuristic semi-dynamic clustering scheme,where the delay of the channel state information feedback via backhaul is considered.We model such system in the framework of networked-POMDP and derive a strategy to reduce the search range for the best cluster configuration.Thereafter,the inter-cluster power control sub-problem is modeled as a set of exact potential games and an effective decentralized algorithm with guaranteed convergence is designed to achieve the Nash Equilibriums.Simulation results and complexity analyses demonstrate that the proposed resource management algorithm not only has much lower computational complexity,but also performs close to the exhaustive search solutions and other existing schemes.