Multistage Stochastic Programming with Recourse As Mathematical Programming in an $L_p $ Space

Abstract

Multistage stochastic programming with recourse has been formulated in terms of a recursive sequence of parameterized, finite-dimensional mathematical programming problems. It has also been formulated as mathematical programming in an $L_p $-space. The two formulations are reconciled by showing that the $L_p $-space, or static, formulation is a restriction of the recursive, or dynamic, formulation and by deriving conditions under which any solution to the static formulation solves the dynamic formulation.