Algorithmic Composition as a Constraint Satisfaction Problem

Throughout the history, musicians have always used music composition rules, which actually are in most cases prohibitions or constraints. The elements of music with which composers work, as well as the possible combinations of those elements, make up a huge finite set of values. The composer’s function is to select, from this big set, a subset of values, which will be arranged into a certain temporal organization, constrained by the music rules, and eventually create an artistic work. From a mathematical point of view, this process is closely related to one field of Artificial Intelligence: Constraint Programming. In this paper the author makes a reflection on the relationship between Music Composition and Constraint Programming. As a concluding idea, the author suggests that Constraint Programming could be a promising technique for the development of computer systems oriented to algorithmic composition. This text follows after the lectures given by professor Dr. Héctor Geffner, in the doctoral course Problem Solving in Artificial Intelligence (2001-2002).