Cultural swarms II: virtual algorithm emergence

Cultural algorithms (CA) (Reynolds 1994) is an evolutionary model derived from the cultural evolution process. CA has two major components, a population components and a belief component. In a previous paper it was show that certain problem solving phases emerged during the optimization process in a dynamic problem solving environment (Reynolds and Saleem 2003). These phases were labeled coarse grained, fine grained and backtracking respectively. I understand how these phases emerged as a result of the interaction of the five knowledge sources influenced individuals in the population component. It was demonstrated in a companion paper (Reynolds 2003) how individuals in an EP population exhibited swarm-like behavior while under the belief space during the search for an optimum in a cones-world environment. In this paper we examine the behavior of the cultural algorithm at the meta-level and demonstrate how, at that level, an algorithmic interaction of knowledge sources emerge. This interaction in terms of best-first search.