CONCLUSIONS
In order to achieve uniform coverage in cellular networks
and required quality of service for multiple users in the MIMO
FDR network, the transmit power at the BS and relays must be
reliably controlled in facing IRI, SI and MUI simultaneously
so that the system can operate uninterruptedly and stably. We
have formulated a power (sum of transmit power of the BS and
relays in the network) minimization problem for the distributed
beamforming design under individual user rate constraints,
where the objective function turns out to be non-strict convex.
Then we have presented a SLIPD method for efficiently
solving this problem, by which the proposed two distributed
beamforming algorithms (Algorithm 1 for the case of single
receive antenna at relays and and Algorithm 2 for the case of
multiple receive antennas at relays) were developed. A globally
optimal solution of the power minimization problem can be
obtained by Algorithm 1, while Algorithm 2, can only yield
a local optimal solution, due to conservative SCA performed
at each iteration, and their convergence can be guaranteed
if the step sizes are below the upper bounds derived via a
rigorous analysis. Moreover, the proposed two algorithms only
require local information exchange between relays, and hence
are highly scalable on one hand. With the suggested early