DISTRIBUTEDTOTALLYASYNCHRONOUS ITERATIVEWATERFILLINGFOR WIDEBAND INTERFERENCECHANNEL WITH TIME/FREQUENCYOFFSET

Thispaperconsiders thecompetitive maximization ofinformation rates intheGaussian frequency-selective interference channel, subject toglobal powerandspectral maskconstraints. We focus on thepractical caseinwhichthetransmission bythedifferent users contains timeandfrequency synchronization offsets. We propose aunified framework based onadistributed algorithm called asynchronous iterative waterfilling algorithm. Inthis algorithm, theusers update their powerspectral density inacompletely distributed and asynchronous way:someusers mayupdate their powerallocation morefrequently thanothers andtheymayevenuseoutdated measurements ofthereceived interference. Moreover theusers arenot required toknowtimeandfrequency offsets. Ourmaincontribution istoprovide aunified setofconvergence conditions forthewhole class ofalgorithms obtained fromtheasynchronous iterative waterfilling algorithm. IndexTerms-Gaussian frequency-selective interference channel, gametheory, Nashequilibrium, totally asynchronous algorithm, iterative waterfilling algorithm.