Development and analysis of evolutionary programming for learning musical notes / Siti Aishah Mohd Noor

• Main Author: Mohd Noor, Siti Aishah
• Format: Thesis
• Language: English
• Published: 2006
• Subjects: Electronic Computers. Computer Science
• Online Access: https://ir.uitm.edu.my/id/eprint/966/2/966.pdf; https://ir.uitm.edu.my/id/eprint/966/
• Institution: Universiti Teknologi Mara

This research is about applying three different types of Evolutionary Programming mutation operators onto musical notes which causing them to learn a small subset of children music notes. Research and study in Evolutionary Programming and its various types of mutations have been implemented. As a result, selected mutation types are obtained in order to perform this project. The musical notation that has been used is "Old McDonald Had A Farm" and it is represented using permutation encoding. The size of population, generations and mutation probability are usually random initialized by user. 48-bit strings of musical notes are randomly generated and it is evaluated by cost functions. The effect of random population size to the performance of each different Evolutionary Programming mutation type has been analyzed. From the experiments, it shows that hybrid mutation is better than one mutation type thus, is able to learn faster for overall populations.