Multicast Scheduling for Relay-Based Heterogeneous Networks Using Rateless Codes

We consider the multicast scheduling problem in the heterogeneous network using a half-duplex relay station (RS). Our goal is to minimize the delay of transmitting a block of packets to users over time-varying channels using rateless codes. Due to half-duplex operation, at each time slot, the RS can choose to either multicast a packet to the users, or fetch a packet from the macro base station. We formulate a fluid relaxation for the optimal decision problem, and reveal that the optimal policy has a threshold-based structure so as to exploit the opportunism of multicast channel: the RS should multicast only when the channel quality is sufficiently "high". We propose an online policy based on the relaxation which does not require the knowledge of channel distribution. When the channel distribution is symmetric across users, we provide a closed-form expression of the asymptotic performance of our policy. For two-user systems, we prove that our scheme is asymptotically optimal. When the users' channels are independent, we derive a performance bound based on water-filling rate allocation which approximates the optimal policy well. Simulation results show that our scheme performs close to theoretical bounds, under correlated as well as independent fading channels.