A generic generic algorithm using phenotype building functions

Abstract

The problem with genetic algorithms is that their use is made difficult by the fact that they are not very reusable because the chromosome representation, genetic operators and fitness function need to be redefined for many problems. In order to use a genetic algorithm you have to think in terms of evolution which might not be easy for a programmer. In this project, we aim to solve these problems by creating a generic chromosome representation which always uses the same genetic algorithm regardless of the problem being solved. The user then directs the genetic algorithm by writing a function, called the 'phenotype building function' or the 'builder function', which maps the generic chromosome into candidate solutions. The fitness function will still have to be provided in order to measure the fitness of a candidate solution.