Optimal packet scheduling over correlated Nakagami-m channels with different diversity-combining techniques

We study two cross-layer optimization problems for M-QAM systems that adapt transmission rate with channel state and buffer occupancy. We formulate both problems as constrained Markov decision process problem and give linear programming technique based solutions. In first problem, our objective is to minimize average transmission power under constraint on average delay and packet dropping probability. We minimize average bit error rate (BER) with average delay and packet dropping probability constraints in second problem. The Nakagami-m fading channel with diversity-combining is described as finite state Markov channel. Simulation results show that the system performances can be improved by adapting rate with buffer state and hence delaying packet in the buffer in addition to employing diversity-combining at the receiver.