In this paper we propose a framework for an energy efficient scheduler for multiuser SC-FDMA with queue state information (QSI) and quality of service (QoS) constraints. Resource allocation is formulated as a two-stage problem where resources are allocated in both time and frequency. The scheduling policy is obtained in two stages for the intra- and inter-user allocations respectively. A near optimal iterative allocation method is used for the inter-user allocation and the intra-user allocation policy is obtained using a constrained Markov decision process framework. Results are presented for the energy performance. © 2012 Elsevier B.V.