Composing Music Using Genetic Algorithms

In this paper, we describe the use of genetic algorithms in composing original music in an evolutionary fashion. We also discuss plans to use the music in smart houses and in music therapy. The music is evolved starting with any given subset of the twelve musical notes. The compositions grow in duration with every generation. In the current version of the software, the fitness of any generation is determined by a human listener. In future versions, the fitness may also be decided by a fitness algorithm that compares the musical pleasantness of the composition with musical landscapes on various genres of music. The output from this software, which is a single melody, can then be used as input to the music rendering software that is being developed by the first-named author, thus generating original music complete with percussion as well as all the harmonies of western music or the special nuances of other genres of music, particularly the gamakams of Indian raagas. Using this software with evolving technologies in the area of robotics, multimedia entertainment systems with real 3-D action can be realized. Making use of evolving sensor and smart house technologies this software can also become the backbone of an intelligent music composition system that can generate original music to suit and/or alter the mood of the occupants of any room in a smart building.