Resource management in QoS-aware wireless cellular networks.

摘要

Emerging broadband wireless networks that support high speed packet data with heterogeneous quality of service (QoS) requirements demand more flexible and efficient use of the scarce spectral resource. Opportunistic scheduling exploits the time-varying, location-dependent channel conditions to achieve multiuser diversity. In this work, we study two types of resource allocation problems in QoS-aware wireless cellular networks. First, we develop a rigorous framework to study opportunistic scheduling in multiuser OFDM systems. We derive optimal opportunistic scheduling policies under three common QoS/fairness constraints for multiuser OFDM systems---temporal fairness, utilitarian fairness, and minimum-performance guarantees. To implement these optimal policies efficiently, we provide a modified Hungarian algorithm and a simple suboptimal algorithm. We then propose a generalized opportunistic scheduling framework that incorporates multiple mixed QoS/fairness constraints, including providing both lower and upper bound constraints.;Next, taking input queues and channel memory into consideration, we reformulate the transmission scheduling problem as a new class of Markov decision processes (MDPs) with fairness constraints. We investigate the throughput maximization and the delay minimization problems in this context. We study two categories of fairness constraints, namely temporal fairness and utilitarian fairness. We consider two criteria: infinite horizon expected total discounted reward and expected average reward. We derive and prove explicit dynamic programming equations for the above constrained MDPs, and characterize optimal scheduling policies based on those equations. An attractive feature of our proposed schemes is that they can easily be extended to fit different objective functions and other fairness measures. Although we only focus on uplink scheduling, the scheme is equally applicable to the downlink case. Furthermore, we develop an efficient approximation method---temporal fair rollout---to reduce the computational cost.