Distributed control of broadcast channels with acknowledgement feedback: Stability and performance

It is well known that Aloha multiaccess schemes, with constant retransmission probability and mere acknowledgement feedback, exhibit bistability for large finite users and are unstable (thruput ?? O) for the infinite user model. We propose an acknowledgement based recursive retransmission procedure which removes bistability for large finite user models and an average success rate of ??e-1 is maintained. For the infinite user Poisson model, the scheme is not unconditionally stable. However, a high level of success rate can be achieved through certain control variations.