In recent times, several activities that lately were considered possible to do only by human work, have been replaced by the use of artificial intelligence. Recent researches have been developed to give computers a creativity ability as main objective. A type of related research is algorithmic composition. Traditionally evolutionary algorithms are successfully used for composing tasks, using a wide variety of fitness functions. This work presents a multiobjective approach that combines Fux's rules, from music theory, and functions that capture melodies' proportions, based on the Zipf's law, to compose monophonic melodies. Besides that, we propose a method that measures the amount of creativity contained in an algorithmic composition method. Experimental results show that the proposed methods have musical quality better than random compositions.
