Time-Frequencv Resource Conversion Based Radio Resource Management And Performance Analvsis For Cellular Networks

Time-frequency resource conversion (TFRC) is a recently proposed network resource allocation strategy via exploiting user behavior information. It can address network congestion and provide quality of experience (QoE) provisioning for users in a wireless broadband mobile network. The key idea is to use radio resources mainly on the traffic/connection to which a user pays attention and thus impacting directly on user's QoE. As such, we can withdraw radio resources from a connection without user focus to generate virtual spectrum hole, which helps a network serve more data traffic or improve user's QoE. For TFRC strategy's better application in practical system, in this thesis we study two problems:TFRC-based centralized resource scheduling and distributed call admission control (CAC).We first study downlink scheduling based on TFRC for an LTE-type cellular network. The service scheduling of interest is formulated as a joint request, channel, and slot allocation problem, which is NP-hard. To solve this problem, we transform the original optimization problem by a usage-based pricing and penalty based adjustment. In order to increase user satisfaction, a new penalty function is proposed for uncompleted requests. Whereby an offline deflation and sequential fixing based polynomial-time algorithm is proposed to solve the problem. Simulation results show the advantage of TFRC strategy in terms of improving both the revenue of a network operator and the request complete ratio of a wireless user. It also shows the effectiveness of the proposed deflation based approach.For practical implementation, we further study TFRC-based online scheduling for on-demand data service. Based on a newly proposed TFRC-oriented online request model, we design two TFRC-enabled online scheduling algorithms. To find the performance gap between TFRC-enabled and TFRC-disabled online algorithms, we propose a new performance metric, TFRC-oriented competitive ratio (TOCR), which is the ratio of the total reward achieved by TFRC-enabled schedule to that by TFRC-disabled counterpart. By considering a single channel with a fixed transmission rate, we derive the upper bound of TOCR for both the new online algorithms. Simulation results verify the analysis and show that TFRC strategy can generate a win-win situation for both the operator and the wireless users.To relieve the impact of a centralized TFRC strategy on a network's overhead, we also study multiuser coupled yet distributed radio resource management approach. Specifically, based on auction theory, we study distributed joint call admission control and radio resource management problem under each user's QoE risk constraint. The proposed scheme can make CAC decision according to the state of both the base station and all users in the cell, which thus can well adapt to network traffic fluctuation and achieve the goal of balancing generating sufficient virtual spectrum holes and providing QoE provisioning for wireless users. Simulation results show that TFRC strategy can well improve the access ratio of requests and total system reward, especially for high priority requests, at a certain cost of low priority requests' access ratio and delay performance.