Personalized music distribution

New media distribution channels have created a strong need for digital media personalization that helps both users and producers get the most of media products. We concentrate on the customized music delivery issue. We propose an approach for constructing a sequence of tracks that satisfies user requirements and at the same time optimally exploits music catalogs. An optimal solution consists of examining all possible tracks enumeration in the database. This is clearly a combinatorial NP-hard problem. We use vector space concepts to formulate the constrained sequence retrieval problem as an integer program. Our experiments demonstrate the power of our approximation to the original problem in reducing the search space and producing valid solutions