Evolving Boolean Networks on Tunable Fitness Landscapes

This paper presents an abstract, tunable model by which to explore aspects of artificial genetic regulatory networks and their design by simulated evolution. The random Boolean network formalism is combined with the NK and NKCS models of fitness landscapes. This enables the systematic study of the interactions between the underlying genetic machinery and elements of the phenotype produced. Previously reported results from the models individually are explored within this context, using both synchronous and asynchronous updating. The evolution of network size is then explored in particular under varying conditions.