Stack algorithms for random multiple access in the presence of asymmetric channel feedback erasures

We propose modifications to the basic free access, limited feedback sensing stack collision resolution algorithm for operation over multiple access networks with ternary empty versus success versus collision (E/S/C) feedback that can improve throughput performance by more than 10%. We also examine via simulation the performance of the proposed stack algorithms in multi-access networks in which each user independently with probability 1-p sees the correct channel feedback, and with probability p can make no decision as to the transmission outcome on the channel (erasure feedback). In the latter case, it is shown that the modified stack algorithms achieve higher throughput than the basic stack algorithm when the erased feedback is treated by the corresponding users as a collision