Probabilistic resource allocation and scheduling for multiuser communication systems

This paper describes how resource allocation and scheduling of users in a multiuser communication system, such as UMTS, can be used to enhance system throughput. More specifically, it suggests a framework based on probability theory for allocating base stations and time slots in the downlink, in order to maximize the average throughput. The decisions are based on knowledge of future channel capacities and the average amount of incoming data. It is demonstrated in simulations that the proposed scheduler achieves significant performance improvements over simple versions of traditional scheduling algorithms.