A Versatile Genetic Algorithm System in Ada: Implementation and Applications
Max Edwards
As is explained and justified in section 3.3., the tasks involved in the development of the VAGAS system, and the research carried out using that system, were executed side by side, and to a large extent the many components of this project provided inspiration for each other. For the sake of clarity however, the work that was done is presented in this report as a series of distinct activities.
Chapter 2 of this report consists of a review of the background theory and applications of Genetic Algorithms, in which GA's are put into context amongst other optimisation methods. In chapter 3, the need for a GA software system is explained, and the design of one is described in terms of the requirements analysis, specification, implementation and testing stages. In chapter 4 the details of research carried out using the system are given, and in chapter 5 the results of the various test and experimental programs are presented. Chapter 6 draws conclusions from these results and points the way forward for future work. A programmer's guide for the VAGAS system, and a complete set of listings, can be found in the appendix.
Whilst care has been taken to explain and illustrate unusual or advanced uses of Ada in the project, a certain amount of familiarity with the language has by necessity been assumed in this report. The complete novice is therefore referred to Barnes' Programming in Ada (Barnes 1989) as a solid introductory text.
As a general genetic algorithms reference text, Goldberg's Genetic Algorithms in Search, Optimization and Machine Learning (Goldberg 1989) is hard to beat in terms of both accessibility and depth, and is frequently referred to in this report.
This report, and a full set of Ada sources for the system described, will shortly be available via anonymous FTP at
ftp.minster.york.ac.uk
(login as anonymous and give e-mail address as password), in directory
/pub/mark/max
For further details about obtaining the system and information regarding future updates, or to leave comments and suggestions, see the World-Wide Web document
http://dcpu1.cs.york.ac.uk:6666/mark/top_ga.html
This site also contains information relating to other genetic algorithm research at the University of York. These facilities are expected to be available from the beginning of April 1995.