On two-stage portfolio allocation problems with affine recourse

In this paper we propose an approach based on affine parameterization of the recourse policy for the solution of multi-stage optimization problems that arise in the context of allocation of financial portfolios over multiple periods. Such problems are typically dealt with using the multi-stage stochastic programming paradigm, which has the drawback of being computationally intractable. Here, we show that imposing an affine structure to the recourse policy results in an explicit and exact problem formulation, which is efficiently solvable by means of interior point methods for convex second order cone programs.