讨论了当今算法作曲这一研究领域中存在的一些主要问题.评述了这一领域所采用的一系列关键技术,包括Markov链、随机过程、基于音乐规则的知识库系统、音乐文法、人工神经网络技术以及遗传算法.得出的结论是,作曲系统可以朝着集多种方法为一体的混合型系统(hybrid system)的方向发展.系统应在音乐创作的各个层面上提供灵活的人机交互手段,以便提高系统的实用性和有效性.
Some problems on Algorithmic Composition are discussed, and a survey for series of key techniques used in this approach is made, which include Markov chains, stochastic process, musical knowledge based system, musical grammar, artificial neural networks, and genetic algorithms. The conclusion is that the development of music composition system should involve a combination of the existing technologies, i.e. the Hybrid System development. In order to make the system become more practical and effective, there should be some flexible human intervenient ways for different level music composition in the system.