论文信息 - How Fitness Structure Affects Subsolution Acquisition in Genetic Programming

How Fitness Structure Affects Subsolution Acquisition in Genetic Programming

We define fitness structure in genetic programming to be the mapping between the subprograms of a program and their respective fitness values. This paper shows how various fitness structures of a problem with independent subsolutions relate to the acquisition of subsolutions. The rate of subsolution acquisition is found to be directly correlated with fitness structure whether that structure is uniform, linear or exponential. An understanding of fitness structure provides partial insight into the complicated relationship between fitness function and the outcome of genetic programming’s search.

D. Goldberg

[1] L. Darrell Whitley,et al. Genetic Operators, the Fitness Landscape and the Traveling Salesman Problem , 1992, PPSN.

[2] John R. Koza,et al. Genetic programming - on the programming of computers by means of natural selection , 1993, Complex adaptive systems.

[3] David E. Goldberg,et al. Genetic Algorithm Difficulty and the Modality of Fitness Landscapes , 1994, FOGA.

[4] Una-May O'Reilly,et al. An analysis of genetic programming , 1995 .

[5] Tim Jones. Evolutionary Algorithms, Fitness Landscapes and Search , 1995 .

[6] Hitoshi Iba,et al. Extending genetic programming with recombinative guidance , 1996 .

[7] Astro Teller,et al. Neural Programming and an Internal Reinforcement Policy , 1996 .

[8] Una-May O’Reilly. Using a distance metric on genetic programs to understand genetic operators , 1997, 1997 IEEE International Conference on Systems, Man, and Cybernetics. Computational Cybernetics and Simulation.

[9] Una-May O'Reilly. The Impact of External Dependency in Genetic Programming Primitives , 1998, Canadian Conference on AI.

[10] David E. Goldberg,et al. Where Does the Good Stuff Go, and Why? How Contextual Semantics Influences Program Structure in Simple Genetic Programming , 1998, EuroGP.