Genetic Algorithms

1831 Words4 Pages

Abstract

Genetic algorithms are a randomized search method based on the biological model of evolution through mating and mutation. In the classic genetic algorithm, problem solutions are encoded into bit strings which are tested for fitness, then the best bit strings are combined to form new solutions using methods which mimic the Darwinian process of "survival of the fittest" and the exchange of DNA which occurs during mating in biological systems. The programming of genetic algorithms involves little more than bit manipulation and scoring the quality of solutions. Genetic algorithms have been applied to problems as diverse as graph partitioning and the automatic creation of programs to match mathematical functions.

Introduction

Genetic algorithms are a randomized search method which "breeds" effective solutions to problems through simulation of Darwinian Evolution. Large numbers of potential solutions are created at random. The solutions which show the most promise are then breed together to produce new solutions which receive most of their 'genetic stock' from the better solutions in the previous generation. This is similar to the "survival of the fittest" shown in biological systems, where the individuals which are best adapted to their environment breed more offspring, resulting in the better adapted genetic material carrying forward into future generations.

The history of genetic algorithms is most commonly traced to Holland's text "Adaptation in Natural and Artificial Systems", published in 1975. Earlier works by Holland and others shows that the concept of genetic algorithms first began to form in the late 1960s [1]. Around that time, Bagley first coined the phrase "genetic algorithm" in his dissertation. Holland...

... middle of paper ...

...Come of Age, David E. Goldberg, Communications of the ACM, March 1994, pp 113-119.

[3] Genetic Algorithms: A Survey, M. Srinivas, Latit M. Patnaik, IEEE Computer, June 1994, pp 17-26.

[4] Genetic Algorithm and Graph Partitioning, Thang Nguyen Bui, Byung Ro Moon, IEEE Transactions on Computers, July 1996, pp 841-855.

[5] http://www.mines.edu/students/d/drferrin/Cool_Beans/GeneMachiene.html

[6] Genetic Programming: A Paradigm for Genetically Breeding Populations of Computer Programs to Solve Problems, John R. Koza, Stanford Technical Report STAN-CS-90-1314, June 1990. Available at ftp://elib.stanford.edu/pub/reports/cs/tr/90/1314/. This material is also discussed in Koza's text "Genetic Programming: On the Programming of Computers by Means of Natural Selection".

[7] http://www.ifh.ee.ethz.ch/~gerber/approx/default.html

[8] http://www.novagenetica.com

Open Document