SPCoding School

Decoders_Library.pdf

A short description of the project

(141.78 KiB) Downloaded 242 times

Despite the fact that Maximal Likelihood (ML) decoders and Nearest Neighbor (NN) decoders are easy to describe, they are in general extremely difficult to be implemented. It the most common
setting, considering a k-dimensional linear code in an n-dimensional vector space (over a finite field) a search algorithm for syndrome decoding involves a list of the size that increases exponnetially with n-k.

Small values of n are suitable for situations with strong constrains at block length and for research proposes, since actually listing decoders is essential for computing many of the important invariants (such as error probability).

The main goal of this project is to create a library of ML and NN decoders for use with software for numerical analysis, such as Mathematica, Maple, Matlab, etc.

Looks interesting. And moreover, could be useful in every-day work.