Search for a command to run...
The current dominate layered networking architecture, in which each layer is designed and operated independently, results in inefficient and inflexible resource use in wireless networks due to the nature of the wireless medium, such as time-varying channel fading, mutual interference, and topology variations. In this thesis, we focus on resource allocation and scheduling in wireless orthogonal frequency division multiplexing (OFDM) networks based on joint physical and medium access control (MAC) layer optimization. To achieve orders of magnitude gains in system performance, we use two major mechanisms in resource management: exploiting the time variance and frequency selectivity of wireless channels through adaptive modulation, coding, as well as packet scheduling and regulating resource allocation through network economics. With the help of utility functions that capture the satisfaction level of users for a given resource assignment, we establish a utility optimization framework for resource allocation in OFDM networks, in which the network utility at the level of applications is maximized subject to the current channel conditions and the modulation and coding techniques employed in the network. Although the nonlinear and combinatorial nature of the cross-layer optimization challenges algorithm development, we propose novel efficient dynamic subcarrier assignment (DSA) and adaptive power allocation (APA) algorithms that are proven to achieve the optimal or near-optimal performance with very low complexity. Based on a holistic design principle, we design max-delay-utility (MDU) scheduling, which senses both channel and queue information. The MDU scheduling can simultaneously improve the spectral efficiency and provide right incentives to ensure that all applications can receive their different required quality of service (QoS). To facilitate the cross-layer design, we also deeply investigate the mechanisms of channel-aware scheduling, such as efficiency, fairness, and stability. First, using extreme value theory, we analyze the impact of multiuser diversity on throughput and packet delay. Second, we reveal a generic relationship between a specific convex utility function and a type of fairness. Third, with rigorous proofs, we provide a method to design cross-layer scheduling algorithms that allow the queueing stability region at the network layer to approach the ergodic capacity region at the physical layer.