Provably Near-Optimal Approximation Algorithms for Operations Management Models

Abstract

Multistage stochastic optimization models arise in various application domains, including central areas of operations management, such as inventory control, supply chain management, and revenue management. Unfortunately, it is usually computationally very hard to find optimal solutions for these fundamental models, and typically even finding good solutions is very challenging, both in theory and practice. In this tutorial, we will survey various relatively new algorithmic and performance analysis techniques that can be applied to some of these models to construct provably near-optimal algorithms, particularly algorithms that admit worst-case performance guarantees. These techniques span ideas from various disciplines, such as optimization, computer science, and stochastic analysis.