Abstract
In this paper, we present an auxiliary function approach to solve the overlap group Lasso problem. Our goal is to solve a more general structure with overlapping groups, which is suitable to be used in cellular automata (CA). The CA were introduced to the algorithmic composition which is based on the development and classification. At the same time, concrete algorithm and mapping from CA to music series are given. Experimental simulations show the effectiveness of our algorithms, and using the auxiliary function approach to solve Lasso with CA is a potentially useful music automatic-generation algorithm.
Acknowledgements
This research was funded by the third society science and humanities program of Northwest Normal University, project number: NWNU- KJCXGC-SK0303-18, and supported by the subject for the postgraduate tutors of universities belonging to Gansu province (1001-27).