Finite and infinite population models of the canonical genetic algorithm are presented, as well as extentions to include local search and permutation-based representations. We illustrate that the p-vector representing the infinite population trajectory also represented the sampling distribution used by the finite population model. The schema theorem and the concept of implicit parallelism are reviewed. The relationship between these models and other forms of evolutionary algorithms will also be explored, as well as their potential for better understanding parallel evolutionary algorithms.

Back to Workshop Schedule

1996-1997 Mathematics in High Performance Computing