Use of metaheuristics for automatic generation of error corrector systems based on convolutional encoding

• Grant number: 19/16997-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: February 01, 2020
• End date: December 31, 2020
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Denis Gustavo Fantinato
• Grantee: Lucas Fernandes Muniz
• Host Institution: Centro de Matemática, Computação e Cognição (CMCC). Universidade Federal do ABC (UFABC). Ministério da Educação (Brasil). Santo André , SP, Brazil

Abstract

Errors in digital data are a common issue faced by computational systems when dealing with tasks such as transmission and storage of information. Recently, in certain contexts, such as approximate computing, a considerably high level of errors might be tolerated in order to achieve lower power consumption. In these cases, an error control method is required. An option is the error control coding, which is able to detect and to fix the information by introducing redundancy in data. Some of the coding schemes used nowadays are Convolutional, LDPC (Low-Density Parity Check) and Turbo coding. However, although high-speed digital systems are very efficient, the performance limit established by Shannon was not achieved yet. Besides, there remains a gap with coding/decoding schemes when dealing with distortions with memory. With this in mind, this research project aims at finding efficient coding/decoding schemes generated by a machine, willing to achieve Shannon's theoretical limit. Particularly, metaheuristics such as the Genetic Algorithm, BRKGA and Ant Colony Optimization will be considered to perform the search for different convolutional coding schemes. (AU)