QoS Analysis of a Scheduling Policy for Heterogeneous Users Employing AMC Jointly with ARQ

This paper analyzes the quality of service (QoS) of scheduling algorithms for heterogeneous users in multi-user (MU) wireless systems that take advantage from a cross-layer design with both adaptive modulation and coding (AMC) and automatic repeat request (ARQ). By developing a general theoretical framework based on a finite-state Markov chain, we analytically evaluate the average delay, the packet-loss rate (PLR) and the throughput of a scheduling algorithm based on the channel condition, the buffer occupancy, and the number of retransmissions, of users belonging to different service classes. The key assumption of our analysis, i.e., the independence of the stationary states of different users, highly reduces the computational complexity while preserving a sufficient accuracy. To present the proposed analysis, we also suggest an effective scheduling policy suitable for users belonging to different service classes, compliant with the WiMAX standard. The good match between analytical and simulated performance validates our theoretical findings, and enables the proposed approach to be used for cross-layer optimization.