Skip to Main Content
In this paper, we propose the automatic composition system based on genetic algorithm and N-gram model. In the proposed system, the tone and the length of sounds are expressed in a form of gene, and some melodies are generated by genetic algorithm. In this system, the fitness of the gene is calculated based on (a) transition of rhythm, (b) transition of tone, (c) transition of tone and length, (d) transition of the number of sounds per bar, (e) similarity of rhythms between bars, (f) rate of unique sounds of scales and (g) difference between two sounds. We carried out a series of computer experiments and confirmed the effectiveness of the proposed system.