Evolutional modelling of music: principles, approaches, problems
In the published article it is mentioned that beginnings and development of all noted biological systems become feasible owing to evolution processes. Hence the robustness, flexibility and diversity of poly-level organization kinds which are inherent in the named systems are the great stimulus for using of natural evolution principles in different areas of science and artistic work. As example some methods typical for biological evolution on a genetic level, such as crossover, mutation and selection, are conformed to musical art; thus, the genome becomes the base of evolution modelling of music. As applied to a new direction of musical art (that was conceived on the boundary of the XX–XXIst centuries) – the evolutionary music composition, irrespective of the work with a material ways, the definition of fitness function which favours to achieve the artistically inviting or even if stylistically correct results is a difficult problem. Different approaches have been used to resolve this problem: from interactive computation and machine learning to application of premeditated strict rules. This article is based on analysis of the advantages and disadvantages of evolutionary modelling of music according to the current stage of technologies and musicology development. The article presents a typological description of methods for construction of fitness function and its key structures.
interactive music, evolution modelling, fitness function, genetic algorithm, genetic programming.