Improving the Simulated Annealing algorithm for source codeword index assignment by using the mechanism of Tabu search algorithm

Abstract

Channel-optimized Index Assignment (IA) of source codewords is the simple but effective approach of improving the error resilience of the communication systems. Although IA is a type of Joint Source Channel Coding (JSCC), it does not intervene with the source codec design. So, in addition to the fact that this method can be used in designing systems effectively, it can be also applied to the existing system. In the past, several methods and algorithms were developed for the IA problem. Among these IA algorithms, Simulated Annealing (SA) algorithm is the most effective and popular algorithm with fast convergence time, wide use in optimization problems in general and issue of IA problem in particular and has been continuously improved. In this paper, we focus on the MSA algorithm, one of the latest improved versions of SA for the IA problem and improved the MSA algorithm by using the mechanism of TS algorithm in order to avoid repeated searches. We also proposed the modified structure Tabu list applied to the new algorithm in order to extend the searching space and to enhance the algorithm performance. The effectiveness of the proposed method is confirmed through experiments.