Optimal Utility-Based Multi-User Scheduling and Low-Complexity Alternatives

Abstract:
In this work the problem of utility-based multiuser scheduling is considered in a fading environment. The goal is to make use of the multi-user diversity while still guaranteeing short-term fairness. The performance of a scheduler is captured by a utility function. Based on the utility function and assuming statistical knowledge of the time-varying channel the optimization problem for the optimal scheduler can be formulated. Based on the optimal scheduler various simplifications are proposed which use estimations of future system states. Simulations show the performance gain compared to state-of-the-art methods. Index Terms — opportunistic scheduling, predictive scheduling, multi-user diversity.