The social fabric approach as an approach to knowledge integration in Cultural Algorithms

Recently there has been increased interest in socially motivated approaches to problem solving. These approaches include particle swarm optimization, ant colony optimization, and cultural algorithms. Each of these approaches is derived from a social system that operates on potentially different scale. In previous work we introduced a toolkit to model optimization problem solving using cultural algorithms. In this paper we extend the influence and integration function in the cultural algorithm toolkit (CAT) by adding a mechanism by which knowledge sources can spread their influence throughout a population. We then compare this enhanced approach with previous approaches using the Cones world optimization landscape. Dejong and Morrison proposed the Cones world as an alternative to traditional benchmark optimization problems in the assessment of optimization algorithms. We demonstrate how the social fabric enhances cultural algorithm performance within this environment relative to earlier system.