Article ID: | iaor20084732 |
Country: | Brazil |
Volume: | 27 |
Issue: | 3 |
Start Page Number: | 427 |
End Page Number: | 456 |
Publication Date: | Sep 2007 |
Journal: | Pesquisa Operacional |
Authors: | Pereira C.A.B., Fernandez P.J., Lauretto M.S., Stern J.M. |
Keywords: | programming: quadratic |
In the financial markets, there is a well established portfolio optimization model called generalized mean–variance model (or generalized Markowitz model). This model considers that a typical investor, while expecting returns to be high, also expects returns to be as certain as possible. In this paper we introduce a new media optimization system based on the mean–variance model, a novel approach in media planning. After presenting the model in its full generality, we discuss possible advantages of the mean–variance paradigm, such as its flexibility in modeling the optimization problem, its ability of dealing with many media performance indices – satisfying most of the media plan needs – and, most important, the property of diversifying the media portfolios in a natural way, without the need to set up ad hoc constraints to enforce diversification.