Social Learning in Cultural Algorithms with Auctions

Cultural Algorithms provide a metaheuristic framework for the application of various solution mechanism represented by Knowledge Sources in the Belief Space to the evolution of complex social networks in the population space. One of the key components in the configuration of a Cultural Algorithm is the knowledge distribution mechanism, and how the influence of a solution approach is spread out among individuals in individual subnetworks. In this paper, a specific set of Knowledge distribution mechanisms based on various Auction models are introduced and their performance compared in complex real‐valued functional landscapes. While capable of generating solutions to the entire range of problem, from fixed to chaotic, they are best utilized for fixed problems containing of low to high variability but begin to lose their edge with increasing periodicity. It is proposed that auction mechanisms will be most likely found in subcultures where problems associated with that culture have reasonably strong signals that can afford the opportunity to make quick and precise decisions.