Title | ||
---|---|---|
Computer aided proofs for rate regions of independent distributed source coding problems |
Abstract | ||
---|---|---|
The rate regions of independent distributed source coding (IDSC) problems, a sub-class of the broader family of multi-source multi-sink networks, are investigated. An IDSC problem consists of multiple sources, multiple encoders, and multiple decoders, where each encoder has access to all sources, and each decoder has access to a certain subset of the encoders and demands a certain subset of the sources. Instead of manually deriving the rate region for a particular problem, computer tools are used to obtain the rate regions for hundreds of nonisomorphic (symmetry-removed) IDSC instances. A method for enumerating all non-isomorphic IDSC instances of a particular size is given. For each non-isomorphic IDSC instance, the Shannon outer bound, superposition coding inner bound, and several achievable inner bounds based on linear codes, are considered and calculated. For all of the hundreds of IDSC instances considered, vector binary inner bounds match the Shannon outer bound, and hence, exact rate regions are proven together with code constructions that achieve them. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1109/NETCOD.2015.7176794 | 2015 International Symposium on Network Coding (NetCod) |
Keywords | Field | DocType |
independent distributed source coding problems rate region,IDSC problem,multisource multisink network,multiple decoder,computer tool,Shannon outer bound,superposition coding inner bound,linear code,vector binary inner bound match | Linear network coding,Source code,Computer science,Binary code,Theoretical computer science,Distributed source coding,Encoder,Shannon–Fano coding,Decoding methods,Variable-length code | Conference |
ISSN | Citations | PageRank |
2374-9660 | 2 | 0.37 |
References | Authors | |
7 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Congduan Li | 1 | 40 | 7.75 |
Steven Weber | 2 | 724 | 53.55 |
John MacLaren Walsh | 3 | 107 | 17.90 |