Near-optimal resource allocation for type-II HARQ based OFDMA networks under rate and power constraints

We address the problem of multiuser power and bandwidth allocation for a general class of OFDMA-based wireless networks that employ a Type-II hybrid automatic repeat request (HARQ) mechanism along with practical modulation and coding schemes (MCSs). This problem is formulated as minimizing the sum transmit power required to satisfy individual goodput constraints without exceeding maximum allowable per-link or pernode transmit power levels. We assume that the resource manager has only statistical knowledge of the channel state information (CSI) of the Rayleigh-distributed fast-fading links of the network. Using a tight approximation of the goodput, we propose an algorithm allowing computation of the corresponding optimal resource allocation. We finally provide an efficient selection of the different MCSs that can be coupled with the proposed resource allocation algorithm to significantly boost its performance.