Capacity of Random Channels with Large Alphabets
Information Theory
2017-11-28 v3 math.IT
Optimization and Control
Abstract
We consider discrete memoryless channels with input alphabet size and output alphabet size , where ceil for some constant . The channel transition matrix consists of entries that, before being normalised, are independent and identically distributed nonnegative random variables and such that . We prove that in the limit as the capacity of such a channel converges to almost surely and in , where denotes the entropy of . We further show that, under slightly different model assumptions, the capacity of these random channels converges to this asymptotic value exponentially in . Finally, we present an application in the context of Bayesian optimal experiment design.
Keywords
Cite
@article{arxiv.1503.04108,
title = {Capacity of Random Channels with Large Alphabets},
author = {Tobias Sutter and David Sutter and John Lygeros},
journal= {arXiv preprint arXiv:1503.04108},
year = {2017}
}
Comments
20 pages, 2 figures, revised version