User Aggregation Models for a Video-on-Demand (VoD) System

Abstract

We consider the scalability issues in the static data segment location problem in information networks and provide integer linear programming (ILP) formulations for user aggregation. We explore three approaches based on k-means clustering, locality-of-reference and biased locality of reference. The second and third approaches are innovatively modeled as discrete optimization problems. The models are tested in IBM ILOG CPLEX Optimization Studio platform and the computational results are discussed in this paper.