Optimal Design of Degree Distribution in LT Codes Based on Evolution Strategy

Abstract

Degree distribution is the most important component that is responsible for the performance of rateless codes. Evolution Strategy approach has been introduced into the degree optimization process along with Monte Carlo simulation evaluation to obtain a better degree distribution. Compared with Robust Soliton Distribution, this approach can obtain better degree distribution data which have a lower mean value of the degree and a lower successful decoding overhead. Simulation results show that degree distribution optimization approach of LT codes with the evolution strategy algorithm is effective.