Evolved Digital Circuits and Genome Complexity
Morten Hartmann, Per Kristian Lehre, Pauline Catriona Haddow · 2006
A major issue with evolutionary computation is scalability. In the field of digital circuit design this fact severely limits the size and complexity of the circuits that can be evolved. Developmental approaches are being suggested as a possible remedy to the scalability issue. Earlier theoretical work indicated that a Kitano mapping develops phenotypes with some form of regularity. Applying this result to the field of evolvable hardware implies that to develop a digital circuit with a developmental mapping, such as the Kitano mapping, places a requirement of regularity on the digital circuit. This issue of regularity is investigated herein, as well as possible encoding schemes. The range and distribution of the complexity of evolved circuits and legal genotypes is measured using Lempel-Ziv compression.