genetic algorithms in noisy environments（遗传算法在嘈杂的环境中）.pdfVIP

Machine Learning 3: 101-120, 1988 © 1988 Kluwer Academic Publishers - Manufactured in The Netherlands Genetic Algorithms in Noisy Environments J. MICHAEL FITZPATRICK (JMF@VUSE.VANDERBILT.EDU) JOHN J. GREFENSTETTE+ (GREFENSTETTE@VUSE.VANDERBILT.EDU) Computer Science Department, Vanderbilt University, Nashville, Tennessee 37235 (Received: December 23, 1987) (Revised: May 17, 1988) Keywords: Learning in noisy environments, genetic algorithms, image registration. Abstract. Genetic algorithms are adaptive search techniques which have been used to learn high-performance knowledge structures in reactive environments that pro- vide information in the form of payoff. In general, payoff can be viewed as a noisy function of the structure being evaluated, and the learning task can be viewed as an optimization problem in a noisy environment. Previous studies have shown that genetic algorithms can perform effectively in the presence of noise. This work ex- plores in detail the tradeoffs between the amount of effort spent on evaluating each structure and the number of structures evaluated during a given iteration of the ge- netic algorithm. Theoretical analysis shows that, in some cases, more efficient search results from less accurate evaluations. Further evidence is provided by a case study in which genetic algorithms are used to obtain good registrations of digital images. 1. Introduction In approaches to machine learning that use genetic algorithms, success de- pends on the ability to search efficiently for high-performance knowledge struc- tures despite the presense of considerable amoun