Maximizing the throughput guarantees in wireless networks under imperfect channel knowledge

In our earlier work, we described an optimization problem and corresponding scheduling algorithm aimed at obtaining maximum throughput guarantees in wireless networks. To further improve the short-term performance, we also proposed two adaptive versions of the optimal algorithm. Results from the simulations showed that the adaptive algorithms perform significantly better than other well-known scheduling algorithms in networks based on Mobile WiMAX, HSDPA, LTE and WINNER I. However, there were several idealistic assumptions in that analysis, the most important of which is that each user estimates its carrier-to-noise ratio (CNR) perfectly, and there is no feedback delay. In practice however, the channel estimation is not perfect, and there is always some delay in the feedback channel. In this paper, we assume that a maximum a posteriori (MAP) predictor is employed for the CNR, so that the system takes the feedback delay and the channel noise into account. We then investigate the effect of imperfect channel prediction and delay on the throughput guarantees promised to all the users in the wireless network. A procedure to reduce the probability of outage in case of imperfect channel prediction is also proposed.