@brazovayeye

A Comparative Analysis of GP

, and . Advances in Genetic Programming 2, chapter 2, MIT Press, Cambridge, MA, USA, (1996)

Abstract

In order to analyze Genetic Programming (GP), this chapter compares it with two alternative adaptive search algorithms, Simulated Annealing (SA) and Stochastic Iterated Hill Climbing (SIHC). SIHC and SA are used to solve program discovery problems posed in the style of GP. In separate versions they employ either GP's crossover operator or a mutation operator. The comparisons in terms of likelihood of success and efficiency show them to be effective. Based upon their success, hybrid versions of GP and hill climbing are designed that improve upon a canonical version of GP. Program discovery practitioners may find it useful to coherently view all the algorithms this chapter considers by using the perspective of evolution.

Links and resources

Tags