Session:

NuE - Neutral Evolution in Evolutionary Computation

Title:

Six Degrees of Separation in Boolean Function Networks with Neutrality

Authors:

Tina Yu

Abstract:

We analyze two Boolean function networks with different degrees of neutrality. The results show that the one with explicit neutrality is a small-world network where each pair of possible solutions has a short distance and most of the possible solutions are highly clustered. These network structural properties owe their existence to the ``short cuts'' introduced by redundant genes in the genotypes. We explain some important small-world network structures, such as clusters, hubs and power law link distribution. These properties have potential to be useful in designing efficient evolutionary algorithms to navigate search in the network.