Abstract:
This work addresses the computational limitations of the Sample Average Approximation (SAA) method in multistage stochastic programming under Markov-dependent data processes. While SAA is effective for static and two-stage stochastic optimization problems, it becomes computationally prohibitive in multistage settings as the number of samples required to obtain a reasonably accurate solution grows exponentially in the time horizon $T$—a phenomenon known as the curse of dimensionality. To overcome this challenge, we propose a novel data-driven approach: the Markov Recombining Scenario Tree (MRST) method, combined with Stochastic Dual Dynamic Programming (SDDP) as a solution framework. Our analysis shows that MRST achieves polynomial sample complexity in $T$, offering an efficient data-driven alternative to SAA. Extensive numerical experiments further validate the effectiveness of MRST, showcasing its potential to mitigate the curse of dimensionality in multistage stochastic programming.
Bio:
Grani A. Hanasusanto is an Associate Professor in Industrial & Enterprise Systems Engineering at the University of Illinois Urbana-Champaign (UIUC). Previously, he was an Assistant Professor at The University of Texas at Austin and a Postdoctoral Scholar at École Polytechnique Fédérale de Lausanne. He holds a PhD in Operations Research from Imperial College London and an MSc in Financial Engineering from the National University of Singapore. Grani’s research focuses on developing tractable solution approaches for decision-making under uncertainty, with applications in operations management, energy systems, finance, machine learning, and data analytics. His work has been published in leading journals, including Operations Research, Mathematical Programming, SIAM Journal on Optimization, Manufacturing & Service Operations Management, Stochastic Systems, and IEEE Transactions on Power Systems. Grani received the NSF CAREER Award in 2018 and was named a Walker Scholar by the UT Walker Department of Mechanical Engineering, recognizing his contributions to research, teaching, and service. He has served as an INFORMS DEI Ambassador and is currently on the INFORMS DEI Community committee as well as the editorial board of Operations Research as an Associate Editor.
