Distributed dynamic channel resource allocation in wireless communication systems

We consider the problem of allocation of bits and power among N channels in a multi-user cellular system. We suppose that no users within the same cell interfere with each other, and that the N channels assigned to the users are randomized such that any given interferer is encountered in only one of the channels. We present distributed algorithms to allocate power and bits so as to maximize a cost function including as elements the bit rate already achieved and the likelihood of dropping calls already in progress. Procedures for rapidly estimating the likelihood of achieving target bit rates in the channel are also presented