Encoding and decoding algorithms for LP-decodable multipermutation codes

Abstract

LP-decodable multipermutation codes are a class of multipermutation codes that can be decoded using linear programming (LP). These codes are defined using linearly constrained multipermutation matrices, which are binary matrices that satisfy particular row sum and column sum constraints. Although generic LP solvers are capable of solving the LP decoding problem, they are not efficient in general because they do not leverage structures of the problem. This motivates us to study efficient decoding algorithms. In this paper, we focus on encoding and decoding algorithms for LP-decodable multipermutation codes. We first describe an algorithm that “ranks” multipermutations. In other words, it maps consecutive integers, one by one, to an ordered list of multipermutations. By leveraging this algorithm, we develop an encoding algorithm for a code proposed by Shieh and Tsai. Regarding decoding algorithms we propose an iterative decoding algorithm based on the alternating direction method of multipliers (ADMM), each iteration of which can be solved efficiently using off-the-shelf techniques. Finally, we study decoding performances of different decoders via simulation.