Optimization Algorithms for Internet Revenue Management

Abstract

Due to the development of new technology in wireless communication the amount of online media usage has been increasing significantly in recent years. As the number of online media users increases, the revenue management from online advertising becomes a complex task. In general, a revenue management system for online advertising system consists of Inference Engine and Ad Server. Inference Engine predicts users’ profiles based on their historical viewing data while Ad Server allocates users’ viewing (impressions) to advertising campaigns based on their target audience. In this paper, models for advertise optimization (Impression Allocation models) that can be implemented at Ad Server are introduced. Impression Allocation models maximize the revenue by optimally allocating users’ impressions to advertising campaigns. Models as well as the proposed algorithms that can be used to solve the models efficiently are provided.