Rate allocation in multiuser cognitive radio systems with successive group decoding

We consider the problem of rate allocation in classical as well as generalized multiuser cognitive radio systems. Each user intends to communicate with its designated receiver and all receivers employ successive group decoders with specified complexity constraints. Rate allocations in the classical case are obtained by using algorithms designed for a fi user Gaussian Interference Channel (GIC). In the generalized cognitive radio system the transmission rates of the primary users are assumed to be pre-determined such that each primary user is decodable at its (primary) receiver in the GIC consisting only of the primary transmitter-receiver pairs. We investigate the problem of selecting an active set of secondary users that are allowed to co-exist with the primary users and allocating rates to them under the constraints that each primary user achieves its pre-determined rate and no primary receiver decodes any secondary user. Each secondary receiver however is free to decode any other user whose codebook it is aware of. The key feature of the rate allocation algorithms we design is that inspite of using distributed and low-complexity ldquogreedyrdquo sub-routines, they can achieve globally optimal solutions.