Author

Andrew Cerni

Date of Award

2013

Document Type

Thesis

Degree Name

Bachelors

Department

Natural Sciences

First Advisor

McDonald, Patrick

Keywords

Genetic Algorithm, Bach, Johann Sebastian, Chorale

Area of Concentration

Natural Sciences

Abstract

One of the goals of natural computing is to develop and use techniques inspired by nature to solve complex problems using a computer. Genetic algorithms represent one such technique. Inspired by the processes of natural selection and evolution, a genetic algorithm seeks to generate solutions to an optimization problem via a process which involves the production of many generations of populations whose individuals represent potential or approximate solutions to the given problem. In this thesis, we employ a genetic algorithm approach to the problem of generating four-part chorales in the traditional Western style. We perform a number of computational experiments and produce successive generations of chorales, and demonstrate that the average and maximum fitness of the candidate pool increases. Additionally, we perform another experiment where the genetic algorithm attempts to reconstruct a chorale composed by J.S. Bach, using distorted versions of the original chorale as a seed population.

Share

COinS