genetic programming


genetic programming

[jə‚ned·ik ′prō‚gram·iŋ] (computer science) evolutionary programming

genetic programming

(programming)(GP) A programming technique which extends thegenetic algorithm to the domain of whole computer programs.In GP, populations of programs are genetically bred to solveproblems. Genetic programming can solve problems of systemidentification, classification, control, robotics,optimisation, game playing, and pattern recognition.

Starting with a primordial ooze of hundreds or thousands ofrandomly created programs composed of functions and terminalsappropriate to the problem, the population is progressivelyevolved over a series of generations by applying theoperations of Darwinian fitness proportionate reproduction andcrossover (sexual recombination).

genetic programming

A type of programming that imitates genetic algorithms, which uses mutation and replication to produce algorithms that represent the "survival of the fittest." While genetic algorithms yield numbers, genetic programs yield ever-improving computer programs. Written in languages such as LISP and Scheme, genetic programming requires the determination of a fitness function, which is a desired output (result). The degree of error in the fitness function determines the quality of the program. For more information, visit www.geneticprogramming.com.