Performance analysis of metaheuristic technique based decoding algorithms for block codes

• Published in: 2017 International Conference on Communication and Signal Processing (ICCSP) pp. 0070 - 0075
• Main Authors: Kolisetty, Harish, Vadlamudi, Sai Vamsi Krishna, Chennam, Vamsi Krishna, Raju, R. Reddy, Reddy, S. Bhargav Anand
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.04.2017
• Subjects: Algorithm design and analysis; Chase-II Algorithm; Decoding; Genetic Algorithm; Genetic algorithms; Genetic Neural Network decoding; Linear Block Codes; Modified Genetic Neural Network decoding; Particle Swarm Optimization; Reliability; Signal processing algorithms; Sociology; Statistics
• DOI: 10.1109/ICCSP.2017.8286562

Metaheuristic algorithms are widely used to obtain optimal solutions for problems where only partial data is available or when computational capacity is limited. This feature of metaheuristic algorithms facilitates their introduction into channel decoding problem which is actually a search for the most likely transmitted codeword from a search space of 2 n codewords in the case of binary codes and q n codewords in the case of q-ary codes. Genetic Algorithm (GA) - a metaheuristic technique-is used here for optimizing the search for the most likely transmitted codeword in channel decoding. Neural Network (NN) - an artificial intelligent algorithm is used for pattern classification in channel decoding. Combining the optimization capability of GA and pattern classification advantage of NN in the decoding process, Genetic-Neural network based Decoding (GND) algorithm with less decoding complexity has been proposed. The use of channel statistical information in decoding process ensures a significant improvement in performance. Exploiting this advantage a metaheuristic technique based Modified GND (Mo-GND) algorithm is proposed in this paper. The proposed Mo-GND algorithm promises better performance compared to all Soft Decision Decoding (SDD) algorithms while trading off complexity. Particle Swarm Optimization (PSO) another metaheuristic technique has an added advantage of faster convergence compared to GA. Applications which cannot afford the complexity trade off as in Mo-GND algorithm can make use of the fast converging decoding scheme based on PSO.