MACHINE LEARNING TECHNIQUES - LASA
MACHINE LEARNING TECHNIQUES - LASA
MACHINE LEARNING TECHNIQUES - LASA
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
177<br />
8 Genetic Algorithms<br />
We conclude these Lecture Notes by covering Genetic Algorithms, a form of stochastic<br />
optimization that can be used to learn parameters of several of the methods seen previously in<br />
these notes. This introduction to Genetic Algorithms is very brief and serves only as a brief<br />
overview of the main concepts.<br />
The term genetic algorithm refers to a model introduced by John Holland in 1975 [Holland, 1975].<br />
Genetic Algorithms are a family of computational models inspired by Darwin’s principle of<br />
selective evolution. The algorithms encode a potential solution to a specific problem on a simple<br />
chromosome-like data structure and apply recombination operators to these structures so as to<br />
preserve critical information.<br />
An implementation of a genetic algorithm begins with a population of typically random<br />
chromosomes. One, then, evaluates these structures and allocates reproductive opportunities in<br />
such a way that those chromosomes, which represent a better solution to the target problem, are<br />
given more chances to reproduce than those chromosomes, which are poorer solutions. The<br />
“goodness” of a solution is typically defined with respect to the current population.<br />
Genetic algorithms are often viewed as optimization tool, although the range of problems to which<br />
genetic algorithms have been applied is quite broad and goes beyond simple function<br />
optimization. They are used to find solutions to complex systems in domains such as:<br />
• Finances: Market predictors, risk evaluators<br />
• Business: Scheduling, time and storage optimization<br />
• Engineering: Dynamics of particles, fluids, etc.<br />
© A.G.Billard 2004 – Last Update March 2011