Sources of books on termination conditions in genetic algorithms

Question

I am currently writing an article on genetic algorithms. I would like to have a small subsection on the termination condition, which determines when the algorithm should stop.

I found this nice site:

http://www.nd.com/genetic/termination.html

which lists some good methods by which the termination condition is determined in genetic algorithms. However, the professor who is responsible for my article is very reluctant to accept random websites as sources.

Can you give some recommendations on books (it would be very good if I could pick them up in the form of an electronic book) that contain more information about these conditions of completion and, possibly, expand the various advantages or disadvantages of the method?

Answer 1

Unfortunately, although this is a very interesting area of ​​research, so far it has paid little attention.

Although the book originally asked the initial question, you might be interested in this published article that discusses some of the completion criteria: “On Terminating Criteria for Genetic Algorithms” (Martín Safe, Jessica Carlellido, Ignacio Ponzoni and Nelida Brignoles) at this URL: http: // www.springerlink.com/content/cgt7635e3lhu3pdn/

You may also be interested in reading several pages of the Ph.D. thesis "Analysis and Improvement of Genetic Algorithms Using Concepts from the Theory of Information" by John Milton, section 3.6 (pages 127 and subsequent), which discusses some conditions for completion. Available at: http://epress.lib.uts.edu.au/scholarly-works/bitstream/handle/2100/1004/02Whole.pdf?sequence=2

Answer 2

Your best bet is probably any good general introductory text on genetic algorithms, which will probably contain a brief description of a possible termination criterion. The only such book I had to transmit was Field Manual on Genetic Programming (GP is very closely related to GA). There is a short paragraph (section 3.5), which refers to the maximum number of generations or the "predicate of success for a specific task." In practice, these two are the most used, with the latter simply meaning stopping whenever a complete solution is found.