A genetic and social computational model for the emergence of skill-based agent specialization

Abstract

There are several methods that lead to the emergence of specialization in agent societies. Two such methods are the Genetic Threshold Model (GTM) and the Social Inhibition Model (SIM). Based on the premises of these models, such as the availability of social networks, or the presence of genetic thresholds, it is difficult to compare results across these models. We present a model that can mimic both these models, while aiming to increase the effect of agent skill on task choice when agents possess different aptitudes for tasks. Using a metric that quantifies the quality of work performed, we are able to see meaningful increases in work quality, but with a side effect of reduced levels of specialization. © 2012 Springer-Verlag.