Weaving the social fabric: The past, present and future of optimization problem solving with cultural algorithms

The purpose of this paper is to investigate the performance of cultural algorithms (CAs) over a complete range of optimization problem complexities, from fixed to chaotic and specifically observing whether there is a given homogeneous agent topology within a culture which can dominate across all complexities.

In order to apply the CA overall complexity classes it was necessary to generalize on its co‐evolutionary nature to keep the variation in the population across all complexities. First, previous CA approaches were reviewed. Based on this the existing implementation was extended to produce a more general one that could be applied across all complexity classes. As a result a new version of the cultural algorithms toolkit, CAT 2.0, was produced, which supported a variety of co‐evolutionary features at both the knowledge and population levels. The system was applied to the solution of a 150 randomly generated problems ranging from simple to chaotic complexity classes.

No homogeneous social fabric tested was dominant over all categories of problem complexity; as the complexity of problems increased so did the complexity of the social fabric that was need to deal with it efficiently. A social fabric that was good for fixed problems might be less adequate for periodic problems, and chaotic ones.

The paper presents experimental evidence that social structure of a cultural system can be related to the frequency and complexity type of the problems that presented to a cultural system.