Modelling of Evolvable Hardware

adaptive and evolvable harware, genotype, FPGA, configuration memory

This paper presents an introductory study to the evolvable hardware design. After basic notions the FPGA structure is described. This type of programmable circuits suits particularly well for the studies of evolution approaches as a fast experimentation tool that surpasses conventional software simulation tools. A method of graph rewriting for the circuit connection matrix of the initial genotype is introduced. Then a standard genetic algorithm can be used to evolve the circuit genotype. The standard GA uses a two-point crossover, adaptive mutation, elite strategy, and real-value coding.

Introductory study to the evolvable hardware design. After basic notions the FPGA structure is described. This type of programmable circuits suits particularly well for the studies of evolution approaches as a fast experimentation tool that surpasses conventional software simulation tools. A method of graph rewriting for the circuit connection matrix of the initial genotype is introduced. Then a standard genetic algorithm can be used to evolve the circuit genotype. The standard GA uses a two-point crossover, adaptive mutation, elite strategy, and real-value coding.