The methods described in this thesis address the problems of both music arranging and tablature generation for the guitar. Arranging is the process by which a piece of music is adapted so that it can be performed on an instrument for which it was not originally written. It is interpreted here as an optimization problem, the goal of which is to establish the most desirable set of notes from the original composition. In the pursuit of this goal, new methods for the automatic generation of guitar tablature were devised. Tablature is a notation system from which the majority of western guitar players read music, and contains information essential to determining the difficulty of a piece of guitar music. Genetic Algorithms are initially employed to solve both the arranging and tablature problems, and an artificial neural network is introduced at a later stage as a faster and more accurate solution to the tablature problem.

Creating Tablature and Arranging Music for Guitar with Genetic Algorithms and Artificial Neural Networks