A new evolvable hardware approach to digital circuits using cultural algorithms

A new evolvable hardware approach using cultural algorithms in order to design digital circuits is presented in this paper. The individuals that represent hardware circuits are designed. The behavior of each individual is evaluated and the best circuits are combined to generate new and better circuits. This evolvable hardware approach supports the collective evolution process through the employments of a population space and a belief space, and extracts domain knowledge of the evolutionary process to decrease computational cost. The implementations of the population space, belief space, and their communications are given in detail. The evolutionary programming technique is used in population space. Experimental results show that the evolvable circuits using the proposed approach are equivalent to or even better than that of human designer and the techniques based on genetic algorithms, therefore the approach proposed in this paper is a viable alternative to the evolving design of digital circuits.