The Selection of Optimal R+D Projects and Policies: a Bayesian approach

Abstract

The selection of an optimal portfolio of R+D projects means maximizing a multicriteria objective function which considers the generally continuous economic variables as well as the discreet variables associated with political decisions. When uncertainty is added, the function of optimizing involves quadratic terms which appear as a result of variances and co-variances. The optimal result will be a Pareto optimum, especially when some of the objectives are opposed to each other.

Uncertainty is managed by using Bayesian statistics to consider expert knowledge in prior estimates. Bayes’ theorem allows us to obtain subsequent estimates as additional evidence appears, which in turn facilitates the monitoring of the optimum portfolio over the course of its evolution as well as the generation of rational and coherent decisions about whether to continue with some projects or to incorporate others. In this work, a general outline of the method is expounded in addition to the methodology for solving the problem of optimization