Low Complexity Utility Based Resource Allocation for 802.16 OFDMA Systems

A utility based resource allocation scheme is investigated for improving system performance of 802.16 OFDMA systems supporting a mix of multi-media traffic types. We show that a utility based approach can improve system capacity over conventional proportional-fair resource allocation schemes through more efficient allocation of only the required resources demanded by the QoS profile of each traffic class. It is also shown that using average channel- aware metrics for resource allocation allows for local linearization of the utility function, which results in linear complexity OFDMA resource allocation schemes. Further, performance with the reduced complexity scheme is similar to a previously known near-optimal algorithm with cubic complexity.