Design and implementation of a genetic algorithm with integer number coding for the evolution of FPGAs in space applications

Abstract

In [1] a form of representation of logic circuits by chains of integer numbers is presented. That type of representation is easy to simulate and to export to FPGA hardware in such a way that, by adding a genetic algorithm (GA), it can be used in an evolutionary process. Because regular GAs utilize binary number coding, one was designed, with all its operators and processes, that uses integer number coding. This evolvable hardware (EH) process was tested with more than 200 hours of runs to determine the effectiveness of the integer coded GA. Results show that, given the proper conditions, the GA is effective in finding solutions that fulfill the required needs of the target system and that this particular EH platform is suitable for applications where fault tolerance capability is required, such as space systems.