Throughput Optimization in High Speed Downlink Packet Access (HSDPA)

摘要

In this paper, we investigate throughput optimizationudin High Speed Downlink Packet Access (HSDPA). Specifically,udwe propose offline and online algorithms for adjustingudthe Channel Quality Indicator (CQI) used by the network toudschedule data transmission. In the offline algorithm, a givenudtarget BLER is achieved by adjusting CQI based on ACK/NAKudhistory. By sweeping through different target BLERs, we canudfind the throughput optimal BLER offline. This algorithm couldudbe used not only to optimize throughput but also to enable fairudresource allocation among mobile users in HSDPA. In the onlineudalgorithm, the CQI offset is adapted using an estimated shortudterm throughput gradient without specifying a target BLER. Anudadaptive stepsize mechanism is proposed to track temporal variationudof the environment. We investigate convergence behaviorudof both algorithms. Simulation results show that the proposedudoffline algorithm can achieve the given target BLER with goodudaccuracy. Both algorithms yield up to 30% HSDPA throughputudimprovement over that with 10% target BLER.