Predictive Resource Allocation for Multicast OFDM Systems

In this paper, we propose a predictive resource allocation algorithm for multicast downlink OFDM systems with imperfect channel state information (CSI). We assume that the CSI is fed back to the base station (BS) through a delayed channel. To avoid the system degradation caused by the outdated CSI, we present a channel prediction scheme at the user side, which takes the delay time into account. Then, under the assumption of the erasure coding, we formulate the optimization problem to maximize the weighted sum rate of the multicast service conditioned on the predicted channel gains, subject to total transmit power and average bit error ratio (BER) constraints. To reduce the time complexity, we propose a two-step suboptimal algorithm that separates subcarrier allocation and power allocation. Simulation results show that the proposed algorithm has much less average BER violation than the algorithm where no prediction is performed; for short feedback delay, the performance of the proposed algorithm is close to the perfect CSI one.