Toward optimizing cauchy matrix for cauchy Reed-Solomon code

The computational costs of Cauchy Reed-Solomon (CRS) encoding operation make a great impact on the performance of its practical applications. The letter concentrates on how to construct a good Cauchy matrix which can lead to an efficient CRS coding scheme. We first formally model the problem by using a binary quadratic programming, then present an approximate method called localized greedy algorithm (LGA) to solve it. Compared with existing work, LGA requires much lower complexities to obtain the same performance of Cauchy matrices.