Retransmission Control Policies in Multiaccess Channels.

摘要

Aloha-type retransmission control policies have been proposed recently by Hajek and Van loon which can be implemented on a random access channel with ternary or binary feedback. They have shown that such schemes achieve a stable throughput of 1/e = .3678 for an infinite-population Poisson arrival model, by using simple first-order recursive retransmission policies. We drive measures of the speed of convergence and steady state accuracy for the local 'model' of these policies. Using these measures, we compare two first-transmission policies, namely IFT and DFT. We extend the policies to cover the Success/Failure binary feedback case, which is not covered in the work cited. Finally, we study the effects of channel errors on the performance of the random access system.