Defining length

In genetic algorithms and genetic programming defining length L(H) is the maximum distance between two defining symbols (that is symbols that have a fixed value as opposed to symbols that can take any value, commonly denoted as # or *) in schema H. In tree GP schemata, L(H) is the number of links in the minimum tree fragment including all the non-= symbols within a schema H.[1]

Example

Schemata "00##0", "1###1", "01###", and "##0##" have defining lengths of 4, 4, 1, and 0, respectively. Lengths are computed by determining the last fixed position and subtracting from it the first fixed position.

In genetic algorithms as the defining length of a solution increases so does the susceptibility of the solution to disruption due to mutation or cross-over.

gollark: There *are* more options than those.
gollark: I mean, you can do it, it would just probably be a bad idea.
gollark: Over here all universities cost the same for in-country students, and they select almost entirely on academic whatever, which is nice.
gollark: Still unlikely to be worth it if you don't really need it for job reasons.
gollark: Okay, so maybe several tens of kilodollars instead?

References

  1. "Foundations of Genetic Programming". UCL UK. Retrieved 13 July 2010.

This article is issued from Wikipedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.