About

Arthur Delarue is an assistant professor and Gary C. Butler Family Faculty Fellow in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech. He completed his Ph.D. in Operations Research at the Massachusetts Institute of Technology in 2021, where he was advised by Dimitris Bertsimas. His educational background also includes a Bachelor of Science in Physics and Mathematics from MIT, earned in 2016. Prior to joining Georgia Tech, he was a postdoctoral fellow at Lyft Rideshare Labs. His primary research focus is on leveraging data, optimization, and machine learning to solve practical problems that matter to society. He is particularly interested in applications of mixed-integer optimization in transportation, machine learning, educational operations, and public policy.

Research topics

• Computer Science
• Artificial Intelligence
• Psychology
• Political Science
• Medicine
• Operations research
• Computer network
• Engineering
• Pedagogy
• Social psychology
• Parallel computing
• Medical emergency
• Economics
• Operations management
• Mathematics education
• Nursing
• Medical education
• Transport engineering

Selected publications

• Learning Pay Strategies with Small Samples in Gig Economy Platforms

  SSRN Electronic Journal · 2026-01-01

  preprintOpen access1st authorCorresponding
  PublisherDOI

• Reinforcement Learning with Combinatorial Actions: An Application to Vehicle Routing

  arXiv (Cornell University) · 2025-01-01 · 5 citations

  articleOpen access1st authorCorresponding

  Value-function-based methods have long played an important role in reinforcement learning. However, finding the best next action given a value function of arbitrary complexity is nontrivial when the action space is too large for enumeration. We develop a framework for value-function-based deep reinforcement learning with a combinatorial action space, in which the action selection problem is explicitly formulated as a mixed-integer optimization problem. As a motivating example, we present an application of this framework to the capacitated vehicle routing problem (CVRP), a combinatorial optimization problem in which a set of locations must be covered by a single vehicle with limited capacity. On each instance, we model an action as the construction of a single route, and consider a deterministic policy which is improved through a simple policy iteration algorithm. Our approach is competitive with other reinforcement learning methods and achieves an average gap of 1.7% with state-of-the-art OR methods on standard library instances of medium size.

  PublisherDOI

• Adaptive optimization for prediction with missing data

  Machine Learning · 2025-03-24 · 7 citations

  articleOpen access

  Abstract When training predictive models on data with missing entries, the most widely used and versatile approach is a pipeline technique where we first impute missing entries and then compute predictions. In this paper, we view prediction with missing data as a two-stage adaptive optimization problem and propose a new class of models, adaptive linear regression models, where the regression coefficients adapt to the set of observed features. We show that some adaptive linear regression models are equivalent to learning an imputation rule and a downstream linear regression model simultaneously instead of sequentially. We leverage this joint-impute-then-regress interpretation to generalize our framework to non-linear models. In settings where data is strongly not missing at random, our methods achieve a 2–10% improvement in out-of-sample accuracy.

  PublisherOA PDFDOI

• Pricing Experiments in Matching Marketplaces under Interference: Designs and Estimators

  ArXiv.org · 2025-02-26

  preprintOpen access1st authorCorresponding

  Interference between treated and untreated units is a source of bias in marketplace experiments. In this paper, we specifically consider pricing interventions, in which a platform seeks to adjust base pricing levels at the marketplace level in order to increase demand. In a matching marketplace, this type of experiment leads to a crucial design question: should the platform match treated and untreated units differently because they paid different prices? We find that standard estimation techniques are biased, but the sign of this bias depends strongly on this design choice. Bias can be reduced by using the ``shadow price estimator'', which relies on the optimal dual solution of the platform's supply-demand matching problem -- especially when the platform chooses to ignore pricing differences at matching time. We validate our findings both theoretically in a fluid limit setting, and numerically in a finite-sample setting.

  PublisherOA PDFDOI

• Algorithmic Precision and Human Decision: A Study of Interactive Optimization for School Schedules

  Management Science · 2025-11-10

  article1st authorCorresponding

  In collaboration with the San Francisco Unified School District (SFUSD), this paper introduces an interactive optimization framework to tackle complex school scheduling challenges. The choice of school start and end times is an optimization challenge, as schedules influence the district’s transportation system, and limiting the associated costs is a computationally difficult combinatorial problem. However, it is also a policy challenge, as transportation costs are far from the only consequence of school schedule changes. Policymakers need time and knowledge to balance these considerations and reach a consensus carefully; past implementations have failed because of policy issues, despite state-of-the-art optimization approaches. We first motivate our approach with a microfoundation model of the interplay between policymakers and researchers, arguing that limiting their dependency is key. Building on these insights, we propose a framework that includes (1) a fast algorithm capable of solving the school schedule problem that compares favorably to the literature and (2) an interactive optimization approach that leverages this speed to allow policymakers to explore a variety of solutions in a transparent and efficient way, facilitating the policy decision-making process. The framework led to the first optimization-driven school start time changes in the United States, updating the schedule of all 133 schools in SFUSD in 2021, with annual transportation savings exceeding $5 million. A comprehensive survey of approximately 27,000 parents and staff in 2022 provides evidence of the approach’s effectiveness. This paper was accepted by Felipe Caro, Special Issue on the Human-Algorithm Connection. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2024.05834 .

  PublisherDOI

• Adaptive Optimization for Prediction with Missing Data

  London Business School Research Online (London Business School) · 2024-02-02

  preprintOpen access

  When training predictive models on data with missing entries, the most widely used and versatile approach is a pipeline technique where we first impute missing entries and then compute predictions. In this paper, we view prediction with missing data as a two-stage adaptive optimization problem and propose a new class of models, adaptive linear regression models, where the regression coefficients adapt to the set of observed features. We show that some adaptive linear regression models are equivalent to learning an imputation rule and a downstream linear regression model simultaneously instead of sequentially. We leverage this joint-impute-then-regress interpretation to generalize our framework to non-linear models. In settings where data is strongly not missing at random, our methods achieve a 2-10% improvement in out-of-sample accuracy.

  PublisherOA PDFDOI

• Algorithmic Precision and Human Decision: A Study of Interactive Optimization for School Schedules

  2024-07-08

  article1st authorCorresponding

  In collaboration with the San Francisco Unified School District (SFUSD), this paper introduces an interactive optimization framework to tackle complex school scheduling challenges. The choice of school start and end times is an optimization challenge, as schedules influence the district's transportation system, and limiting the associated costs is a computationally difficult combinatorial problem. However, it is also a policy challenge, as transportation costs are far from the only consequence of school schedule changes. Policymakers need time and knowledge to balance these considerations and reach a consensus carefully; past implementations have failed because of policy issues despite state-of-the-art optimization approaches.

  PublisherDOI

• Reducing Marketplace Interference Bias via Shadow Prices

  Management Science · 2024-11-29 · 2 citations

  article

  Marketplace companies rely heavily on experimentation when making changes to the design or operation of their platforms. The workhorse of experimentation is the randomized controlled trial (RCT), or A/B test, in which users are randomly assigned to treatment or control groups. However, marketplace interference causes the stable unit treatment value assumption to be violated, leading to bias in the standard RCT metric. In this work, we propose techniques for platforms to run standard RCTs and still obtain meaningful estimates despite the presence of marketplace interference. We specifically consider a generalized matching setting, in which the platform explicitly matches supply with demand via a linear programming algorithm. Our first proposal is for the platform to estimate the value of global treatment and global control via optimization. We prove that this approach is unbiased in the fluid limit. Our second proposal is to compare the average shadow price of the treatment and control groups rather than the total value accrued by each group. We prove that this technique corresponds to the correct first order approximation (in a Taylor series sense) of the value function of interest even in a finite-size system. We then use this result to prove that, under reasonable assumptions, our estimator is less biased than the RCT estimator. At the heart of our result is the idea that it is relatively easy to model interference in matching-driven marketplaces because, in such markets, the platform mediates the spillover. This paper was accepted by Itai Ashlagi, revenue management and market analytics. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.01881 .

  PublisherDOI

• Reducing Marketplace Interference Bias Via Shadow Prices

  2023-07-07 · 9 citations

  article

  Marketplace companies rely heavily on experimentation when making changes to the design or operation of their platforms. A fundamental challenge in marketplace experimentation is dealing with interference. For instance, consider a ride-hailing platform experimenting with a demand-side price discount. The platform performs a randomized control trial (RCT), or A/B test, where some demand units are offered the discounted price while others are offered the undiscounted price. Because the treated units are more likely to book rides as a result of the discount, they reduce the total supply available to all demand-side units, including the control units. This interference between treatment and control units causes the Stable Unit Treatment Value Assumption (SUTVA) to fail, and consequently induces bias in the standard estimator used to evaluate the value generated by the treatment.

  PublisherDOI

• Algorithmic Precision and Human Decision: A Study of Interactive Optimization for School Schedules

  SSRN Electronic Journal · 2023-01-01 · 1 citations

  articleOpen access1st authorCorresponding
  PublisherDOI

Frequent coauthors

• Dimitris Bertsimas

  13 shared

• Sébastien Martin

  Northwestern University

  8 shared

• Guillaume Goudot

  Hôpital Européen Georges-Pompidou

  7 shared

• Lina Khider

  Inserm

  7 shared

• Tristan Mirault

  Assistance Publique – Hôpitaux de Paris

  7 shared

• Emmanuel Messas

  7 shared

• Jean Pauphilet

  London Business School

  6 shared

• Nassim Mohamedi

  Hôpital Européen Georges-Pompidou

  5 shared

Education

• Ph.D., Operations Research

  Massachusetts Institute of Technology

  2021

• S.B., Physics/Mathematics

  Massachusetts Institute of Technology

  2016