Boolean even-3-parity function synthesis by means of arbitrarry evolutionary algorithms - Comparative study

Abstract

This contribution is introducing novelty method which allows to solve various problems from symbolic regression domain. An idea of symbolic regression was firstly proposed by J. R. Koza in so called genetic programming and by C. Ryan in grammatical e volution. In this contribution main principles of analytic programming are described and explained, including some demonstration of its ability to synthesise suitable solutions called programs as well as in genetic programming or grammatical evolution. A comparative study was also carried out. This comparative study is concerned to Boolean even-3-parity problems from Koza's genetic programming which are solved by means of analytical programming applied by means of four evolutionary algorithms: simulated annealing, genetic algorithm, differential evolution and self-organizing migrating algorithm.