Decoding algebraic-geometric codes by solving a key equation

From MaRDI portal





[For the entire collection see Zbl 0745.00062.] This paper investigates a series of preprints by S. C. Porter, where Berlekamp's decoding algorithm is generalized to the case of a small class of Algebraic-Geometric Codes. The main step of this algorithm is to solve a so-called ``key-equation. Here, it is shown how to generalize this concept to all Algebraic- Geometric Codes and moreover, how to solve the key equation by simple linear algebra operations. Two observations on Porter's methods have motivated the paper: 1. The operations done by Porters algorithm at the so-called ``resultant-matrix, may be considered as a Gaussian algorithm, applied to the transposed matrix. 2. The key equation may be reviewed as linear. The result is a decoding algorithm of complexity order \(O(n^ 3)\), that corrects up to \(\left\lfloor{1\over 2}(d^*-1-g)\right\rfloor\) errors, exactly as the well-known algorithm does. (\(d^*\) is the designed minimum distance of the code, \(n\) its length, and \(g\) the genus of the underlying curve). Moreover, it is shown that both algorithms are equivalent is some reasonable manner.











This page was built for publication: Decoding algebraic-geometric codes by solving a key equation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4027636)