Computer Science – Information Theory
Scientific paper
2011-09-08
Computer Science
Information Theory
23 pages, new and revised version, sumitted to IEEE Trans. on Information Thory, Sept.2011
Scientific paper
In this paper (a) after detailed investigation of the previous equivalent rate regions for general interference channel, i.e., the Han-Kobayashi (HK) and the Chong-Motani-Garg (CMG) regions, we define modified CMG region the equivalency of which with the HK region is readily seen; (b) we make two novel changes in the HK coding. First, we allow the input auxiliary random variables to be correlated and, second, exploit the powerful technique of random binning instead of the HK -CMG superposition coding, thereby establishing a new rate region for general interference channel, as an improved version of the HK region; (c) we make a novel change in the CMG coding by allowing the message variables to be correlated and obtain an equivalent form for our new region in (b), as an improved version of the CMG region. Then, (d) in order to exactly demarcate the regions, by considering their different easily comparable versions, we compare our region to the HK and CMG regions. Specifically, using a simple dependency structure for the correlated auxiliary random variables, based on the Wyner and Gacs-Korner common information between dependent variables, we show that the HK and the CMG regions are special cases of our new region. Keywords. Interference channel, Correlated auxiliary random variables, Common information, Superposition coding, Binning scheme
No associations
LandOfFree
A New Rate Region for General Interference Channel (Improved HK Region) does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with A New Rate Region for General Interference Channel (Improved HK Region), we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and A New Rate Region for General Interference Channel (Improved HK Region) will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-475005