Throughput Maximization in Wireless Systems Using an Adaptive Cross Layer Scheduling

Due to the rapid growth of the Internet and other IP services, there has been a change in the design of recent generation wireless systems to support a variety of data applications from non-real time background traffic to real time streaming video. This has placed an enormous strain on the already tight capacity of wireless systems. Many improvements have been made at the link layer to facilitate diverse QoS support. Further it is necessary to improve packet data admission, scheduling and policing techniques to maximize capacity and user satisfaction. Of the three, efficient scheduling has the greatest impact on increased system capacity and user satisfaction. A number of algorithms have been proposed for scheduling data services. The majority of them focus on either minimizing packet delay or maximizing user throughput. In this paper, a new adaptive cross-layer scheduler design is proposed, which can outperform with respect to both packet delay and user throughput by minimizing a prescribed cost function which includes the current channel and delay states of the packet in the queue.