论文信息 - Eugenic Evolution Utilizing A Domain Model

Eugenic Evolution Utilizing A Domain Model

In this paper we introduce The Eugenic Algorithm with Modeling (TEAM), an evolutionary search algorithm that employs statistical analysis to promote construction of high-fitness chromosomes. A model of gene/fitness correlations is automatically generated to direct the construction process. When applied to the combinatorial optimization problems of finding a maximally weighted cut in a graph and minimizing the two-dimensional Rosenbrock function, TEAM performs well compared to other evolutionary algorithms at evolving high-fitness solutions.

[1] John H. Holland,et al. Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[2] William H. Press,et al. Numerical recipes in C (2nd ed.): the art of scientific computing , 1992 .

[3] Yuval Davidor,et al. Epistasis Variance: A Viewpoint on GA-Hardness , 1990, FOGA.

[4] S. Kirkpatrick,et al. Infinite-ranged models of spin-glasses , 1978 .

[5] Shumeet Baluja,et al. A Method for Integrating Genetic Search Based Function Optimization and Competitive Learning , 1994 .

[6] Kenneth Alan De Jong,et al. An analysis of the behavior of a class of genetic adaptive systems. , 1975 .

[7] Gilbert Syswerda,et al. Simulated Crossover in Genetic Algorithms , 1992, FOGA.

[8] John W Prior. Eugenic Evolution for Combinatorial Optimization , 1998 .

[9] William H. Press,et al. Numerical Recipes in C, 2nd Edition , 1992 .

[10] Risto Miikkulainen,et al. Eugenic Neuro-Evolution for Reinforcement Learning , 2000, GECCO.

[11] William H. Press,et al. Numerical recipes in C , 2002 .

[12] John H. Holland,et al. Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[13] W. J. DeCoursey,et al. Introduction: Probability and Statistics , 2003 .