Convergent cutting-plane and partial-sampling algorithm for multistage stochastic linear programs with recourse

We propose an algorithm for multistage stochastic linear programs with reco urse where random quantities in different stages are independent. The algor ithm approximates successively expected recourse functions by building up v alid cutting planes to support these functions from below. In each iteratio n, for the expected recourse function in each stage, one cutting plane is g enerated using the dual extreme points of the next-stage problem that have been found so far. We prove that the algorithm is convergent with probabili ty one.