A decomposition algorithm for two-stage stochastic programs with approximate rotational invariance

TL;DR

This paper introduces an approximation algorithm for two-stage stochastic programs leveraging approximate rotational invariance of the technology matrix, offering a comparison with the traditional L-shaped decomposition method.

Contribution

It presents a novel decomposition algorithm based on approximate rotational invariance, expanding the toolkit for solving two-stage stochastic programs.

Findings

The proposed method approximates the optimal objective value effectively.

Comparison shows advantages over L-shaped decomposition in certain scenarios.

The approach broadens applicability under approximate rotational invariance assumptions.

Abstract

We propose an algorithm of approximating the optimal objective value of a two-stage stochastic program under an assumption of {\it approximate rotational invariance} of the technology matrix, and compare the method with the L-shaped decomposition.