A software framework for genetic programming

Abstract: Genetic Programming (GP) is an automatic programming methodology usingmechanisms inspired by biological evolution. Genetic Programming has beenapplied successfully for a great number of different purposes, and itsviability grows with the increase of computing capabilities. The aim of thisthesis was to craft a software framework that facilitated the easy creationof GP applications, using lessons learned from research in the field toprovide efficient evolution. The focus was the creation of algorithms in acommon programming language, C++. The framework was implemented with a typeof parse tree at its core, and it is put to use for the purpose of findingmathematical expressions. The system is shown to work in practice, but itsaptness in producing generic C++-algorithms remains unclear.