About

Georgia Perakis is the William F. Pounds Professor of Management and a Professor of Operations Management, Operations Research, and Statistics at the MIT Sloan School of Management. She has been on the faculty at MIT Sloan since July 1998. Her teaching spans undergraduate, MSc, PhD, MBA, and EMBA programs, and she has received multiple awards for excellence in teaching, including the Graduate Student Council Teaching Award, the Jamieson Prize, and the Teacher of the Year award in 2017. Her research focuses on analytics and artificial intelligence, particularly at the intersection of optimization and machine learning, with applications in pricing, revenue management, supply chains, healthcare, and energy. She investigates the theory and practice of analytics, aiming to solve complex and practical problems in various domains. Perakis has published extensively in leading journals such as Operations Research, Management Science, and Mathematical Programming, and has received numerous honors including the NSF CAREER Award, the PECASE Award, and election as an INFORMS Fellow and Distinguished MSOM Fellow. She has been recognized for her leadership and innovation in supply chain management and operations research, receiving awards such as the INFORMS MSOM Distinguished Service Award and the Salzburg Medallion from Syracuse University. Perakis has also served in leadership roles at MIT Sloan, including co-director of the Operations Research Center and Associate Dean for SERC. She is currently the editor-in-chief of the M&SOM journal and has held editorial positions in other prominent journals. She holds a BS in mathematics from the University of Athens and an MS and PhD in applied mathematics from Brown University. She is passionate about supervising students, having graduated 30 PhD and 59 masters students, and has been recognized for inspiring student achievement. Her work has significantly contributed to advancing analytics, optimization, and machine learning in operational decision-making across various industries.

Research topics

• Computer science
• Mathematical optimization
• Economics
• Business
• Microeconomics

Selected publications

• OM Forum—Supply Chain Management in the AI Era: A Vision Statement from the Operations Management Community

  Manufacturing & Service Operations Management · 2026-03-26

  article

  Problem definition: Artificial intelligence (AI) is rapidly transforming the research and practice of supply chain management. Yet its impact depends on how effectively it is integrated with the theories, methods, and fundamental principles of operations management (OM), which must also evolve to account for the informational, incentive, and institutional changes brought by AI. The OM community has an important role and responsibility to lead in shaping not only how AI transforms supply chains but also how the supply chains that enable AI are designed to be sustainable, resilient, and equitable. Methodology/results: This vision statement organizes the discussion around five layers of the interaction between AI and supply chain management: intelligence, execution, strategy, human, and infrastructure. It synthesizes recent research and industry practice to show how AI enhances forecasting, planning, decision making, risk management, and human–machine collaboration and also examines the supply chains that support AI. Finally, it highlights persistent challenges in data quality, model integration, governance, and workforce adaptation. Managerial implications: Realizing AI’s promise in supply chain management requires reliable data and infrastructure, integration of learning and optimization, transparent and explainable decision systems, and a long-term commitment to human–AI collaboration. Together, these elements form the foundation for resilient, adaptive, and trustworthy supply chains in the AI era.

  PublisherDOI

• OM Forum—Supply Chain Management in the AI Era: A Vision Statement from the Operations Management Community

  UNC Libraries · 2026-04-09

  articleOpen access

  Problem definition: Artificial intelligence (AI) is rapidly transforming the research and practice of supply chain management. Yet its impact depends on how effectively it is integrated with the theories, methods, and fundamental principles of operations management (OM), which must also evolve to account for the informational, incentive, and institutional changes brought by AI. The OM community has an important role and responsibility to lead in shaping not only how AI transforms supply chains but also how the supply chains that enable AI are designed to be sustainable, resilient, and equitable. Methodology/results: This vision statement organizes the discussion around five layers of the interaction between AI and supply chain management: intelligence, execution, strategy, human, and infrastructure. It synthesizes recent research and industry practice to show how AI enhances forecasting, planning, decision making, risk management, and human–machine collaboration and also examines the supply chains that support AI. Finally, it highlights persistent challenges in data quality, model integration, governance, and workforce adaptation. Managerial implications: Realizing AI’s promise in supply chain management requires reliable data and infrastructure, integration of learning and optimization, transparent and explainable decision systems, and a long-term commitment to human–AI collaboration. Together, these elements form the foundation for resilient, adaptive, and trustworthy supply chains in the AI era.

  PublisherDOI

• Designing Inclusive Offerings

  Management Science · 2025-11-26

  articleSenior author

  The trend toward inclusivity in product and service design—ensuring equal access to, or benefit from, products and services—has gained significant attention in industries such as beauty, wellness, and fashion. However, to our knowledge, there is currently no rigorous definition of inclusivity in the product line design or assortment optimization literature. This paper addresses the need for rigorous models, frameworks, and methods to create inclusive product lines and to measure the inclusivity of existing ones. We introduce the first formal definition of inclusive offerings and propose two continuous measures of inclusivity. We then formulate the inclusive offering design problem, which aims to select a minimum-cardinality set of offerings from a continuous feature space to ensure suitability for all users. Despite its NP-hardness, we present an optimal algorithm for realistic settings and one-dimensional feature spaces. Additionally, we explore the inclusivity resulting from offerings designed to maximize profit, revealing a nonmonotone relationship between inclusivity and the diversity of the user population. Our methods and metrics are applied to the foundation lines of eleven leading makeup brands, demonstrating their practical relevance and impact to the inclusive beauty movement. This paper was accepted by Victor Martínez-de-Albéniz, operations management. Supplemental Material: The online appendices and data files are available at https://doi.org/10.1287/mnsc.2024.06742 .

  PublisherDOI

• CoRe: Coherency Regularization for Hierarchical Time Series

  ArXiv.org · 2025-02-21

  preprintOpen access

  Hierarchical time series forecasting presents unique challenges, particularly when dealing with noisy data that may not perfectly adhere to aggregation constraints. This paper introduces a novel approach to soft coherency in hierarchical time series forecasting using neural networks. We present a network coherency regularization method, which we denote as CoRe (Coherency Regularization), a technique that trains neural networks to produce forecasts that are inherently coherent across hierarchies, without strictly enforcing aggregation constraints. Our method offers several key advantages. (1) It provides theoretical guarantees on the coherency of forecasts, even for out-of-sample data. (2) It is adaptable to scenarios where data may contain errors or missing values, making it more robust than strict coherency methods. (3) It can be easily integrated into existing neural network architectures for time series forecasting. We demonstrate the effectiveness of our approach on multiple benchmark datasets, comparing it against state-of-the-art methods in both coherent and noisy data scenarios. Additionally, our method can be used within existing generative probabilistic forecasting frameworks to generate coherent probabilistic forecasts. Our results show improved generalization and forecast accuracy, particularly in the presence of data inconsistencies. On a variety of datasets, including both strictly hierarchically coherent and noisy data, our training method has either equal or better accuracy at all levels of the hierarchy while being strictly more coherent out-of-sample than existing soft-coherency methods.

  PublisherOA PDFDOI

• Reducing Food Waste through a Reservation Scheme

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access
  PublisherDOI

• Efficient End-to-End Learning for Decision-Making: A Meta-Optimization Approach

  ArXiv.org · 2025-05-16

  preprintOpen access

  End-to-end learning has become a widely applicable and studied problem in training predictive ML models to be aware of their impact on downstream decision-making tasks. These end-to-end models often outperform traditional methods that separate training from the optimization and only myopically focus on prediction error. However, the computational complexity of end-to-end frameworks poses a significant challenge, particularly for large-scale problems. While training an ML model using gradient descent, each time we need to compute a gradient we must solve an expensive optimization problem. We present a meta-optimization method that learns efficient algorithms to approximate optimization problems, dramatically reducing computational overhead of solving the decision problem in general, an aspect we leverage in the training within the end-to-end framework. Our approach introduces a neural network architecture that near-optimally solves optimization problems while ensuring feasibility constraints through alternate projections. We prove exponential convergence, approximation guarantees, and generalization bounds for our learning method. This method offers superior computational efficiency, producing high-quality approximations faster and scaling better with problem size compared to existing techniques. Our approach applies to a wide range of optimization problems including deterministic, single-stage as well as two-stage stochastic optimization problems. We illustrate how our proposed method applies to (1) an electricity generation problem using real data from an electricity routing company coordinating the movement of electricity throughout 13 states, (2) a shortest path problem with a computer vision task of predicting edge costs from terrain maps, (3) a two-stage multi-warehouse cross-fulfillment newsvendor problem, as well as a variety of other newsvendor-like problems.

  PublisherOA PDFDOI

• Optimal Interventions for Increasing Healthy Food Consumption Among Low-Income Populations

  Management Science · 2025-09-30

  articleSenior author

  More than $60 billion per year in the United States is spent on policies aimed to increase fruit and vegetable (FV) consumption among low-income households. Many of these policy interventions are either monetary (e.g., financial incentives) or education related. The goal of this paper is to improve the performance of these interventions through a more strategic and personalized allocation of funds. This paper introduces a consumer behavioral model for grocery shopping decisions, which is nested into the policymaker’s upper-level optimization problem. The policymaker’s goal is to ensure that the FV spending of all consumers in a given population exceeds a specified threshold by utilizing a small strategic set of different intervention bundles—combinations of monetary and education-related interventions. Although an exact solution to the upper-level problem is intractable, we provide an analytical upper bound on the number of intervention bundles needed to achieve the policymaker’s goal, as well as a method for constructing these intervention bundles and assigning them to individuals based on their characteristics. We demonstrate the practicality of the model and approach using the low-income households in the U.S. Department of Agriculture’s FoodAPS data set. This paper was accepted by Jayashankar Swaminathan, operations management. Funding: This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under [Grant 1745302]. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.02324 .

  PublisherDOI

• Decision-Focused AI in Supply Chains

  Springer series in supply chain management · 2025-09-25

  book-chapter1st authorCorresponding
  PublisherDOI

• The Role of Driver Behavior and Interpretability in the Vehicle-to-Grid Market

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access
  PublisherDOI

• Tight mixed-integer optimization formulations for prescriptive trees

  Machine Learning · 2025-05-29 · 1 citations

  articleOpen accessSenior author

  Abstract We focus on modeling the relationship between an input feature vector and the predicted outcome of a trained decision tree using mixed-integer optimization. This can be used in many practical applications where a decision tree or a tree ensemble is incorporated into an optimization problem to model the predicted outcomes of a decision. We propose novel tight mixed-integer optimization formulations for this problem. Existing formulations can be shown to have linear relaxations that have fractional extreme points, even for the simple case of modeling a single decision tree or a very large number of constraints, which leads to slow solve times in practice. A formulation we propose, based on a projected union of polyhedra approach, is ideal (i.e., the extreme points of the linear relaxation are integer when required) for a single decision tree. Although the formulation is generally not ideal for tree ensembles, it generally has fewer extreme points, leading to a faster time to solve. We also study formulations with a binary representation of the feature vector and present multiple approaches to tighten existing formulations. We show that fractional extreme points are removed when multiple splits are on the same feature. At an extreme, we prove that this results in an ideal formulation for a tree ensemble modeling a one-dimensional feature vector. Building on this result, we also show that these additional constraints result in significantly tighter linear relaxations when the feature vector is low dimensional.

  PublisherOA PDFDOI

Recent grants

• Price of Anarchy and Its Applications

  NSF · $260k · 2008–2012

• Tractable Markdown Optimization for an E-tailer

  NSF · $235k · 2012–2017

• Alleviating Travel Delay Uncertainties in Traffic Assignment and Traffic Equilibrium

  NSF · $315k · 2008–2012

• A Data-Driven and Real-time Approach to Personalized Bundle Recommendation and Pricing; from Theory to Practice

  NSF · $395k · 2016–2020

Frequent coauthors

• Maxime C. Cohen

  35 shared

• Hongqiao Chen

  Nanjing University

  26 shared

• Ming Hu

  26 shared

• Pavithra Harsha

  22 shared

• Lennart Baardman

  Ross School

  20 shared

• Divya Singhvi

  20 shared

• Retsef Levi

  18 shared

• Anna Papush

  IBM Research - Thomas J. Watson Research Center

  18 shared

Labs

• Food Supply Chain Analytics and Sensing InitiativePI

Awards & honors

• Manufacturing and Service Operations Management Distinguishe…
• Salzburg Medallion from the Whitman School of Management at…
• INFORMS Best Paper of the Service Science Section Cluster Aw…
• INFORMS honors (2018)