About

Alex Estes is an Assistant Professor at the Robert H. Smith School of Business. He holds a PhD from the University of Maryland. His research involves exploring current and future paths in collaborative research, with a focus on initiatives involving Amazon Research and AWS. His work emphasizes research collaborations and the development of innovative programs in the field of business.

Research topics

• Computer Science
• Mathematics
• Mathematical optimization
• Machine Learning
• Operations research
• Data Mining
• Microeconomics
• Engineering
• Statistics
• Algorithm
• Economics

Selected publications

• Balancing adaptability and predictability: K-revision multistage stochastic programming

  arXiv (Cornell University) · 2026-01-17

  preprintOpen access

  A standard assumption in multistage stochastic programming is that decisions are made after observing the uncertainty from the prior stage. The resulting solutions can be difficult to implement in practice, as they leave practitioners ill-prepared for future stages. To provide better foresight, we introduce the K-revision approach. This new framework requires plans to be specified in advance. To maintain flexibility, we allow plans to be revised a maximum of K times as new information becomes available. We analyze the complexity of K-revision problems, showing NP-hardness even in a simple setting. We examine, both theoretically and computationally, the impact of the K-revision approach on the objective compared with classical multistage stochastic programming models and the partially adaptive approach introduced in [1, 2]. We develop two MIP formulations, one directly from our definition and the other based on a combinatorial characterization. We analyze the tightness of these formulations and propose several methods to strengthen them. Computational experiments on synthetic problems and practical applications demonstrate that our approach is both computationally tractable and effective in reaching near-optimal performance while increasing the predictability of the solutions produced.

  PublisherDOI

• Balancing adaptability and predictability: K-revision multistage stochastic programming

  ArXiv.org · 2026-01-17

  articleOpen access

  A standard assumption in multistage stochastic programming is that decisions are made after observing the uncertainty from the prior stage. The resulting solutions can be difficult to implement in practice, as they leave practitioners ill-prepared for future stages. To provide better foresight, we introduce the K-revision approach. This new framework requires plans to be specified in advance. To maintain flexibility, we allow plans to be revised a maximum of K times as new information becomes available. We analyze the complexity of K-revision problems, showing NP-hardness even in a simple setting. We examine, both theoretically and computationally, the impact of the K-revision approach on the objective compared with classical multistage stochastic programming models and the partially adaptive approach introduced in [1, 2]. We develop two MIP formulations, one directly from our definition and the other based on a combinatorial characterization. We analyze the tightness of these formulations and propose several methods to strengthen them. Computational experiments on synthetic problems and practical applications demonstrate that our approach is both computationally tractable and effective in reaching near-optimal performance while increasing the predictability of the solutions produced.

  PublisherOA PDF

• Nonmonetary Allocations Under Congestion

  SSRN Electronic Journal · 2023-01-01

  articleOpen access1st authorCorresponding
  PublisherDOI

• Smart Predict-then-Optimize for Two-Stage Linear Programs with Side Information

  INFORMS Journal on Optimization · 2023 · 12 citations

  1st authorCorresponding
  • Computer Science
  • Data Mining
  • Computer Science

  We study two-stage linear programs with uncertainty in the right-hand side in which the uncertain parameters of the problem are correlated with a variable called the side information, which is observed before an action is made. We propose an approach in which a linear regression model is used to provide a point prediction for the uncertain parameters of the problem. We use an approach called smart predict-then-optimize. Rather than minimizing a typical loss function for regression, such as squared error, we approximately minimize the objective value of the resulting solutions to the optimization problem. We conduct computational tests that compare our method with other approaches for optimization problems with side information. The results indicate that our method can provide better objective values in situations where the true model is reasonably close to a linear model. Although the procedure we propose requires a longer time for fitting than existing methods, it requires less time to produce a decision for each given observation of the side information. Supplemental Material: The e-companion is available at https://doi.org/10.1287/ijoo.2023.0088 .

  PublisherDOI

• Slow Rates of Convergence in Optimization with Side Information

  SSRN Electronic Journal · 2021-01-01 · 1 citations

  articleOpen access1st authorCorresponding
  PublisherDOI

• Greedy Switching Functions on Time Scales

  SSRN Electronic Journal · 2021-01-01

  articleOpen access1st authorCorresponding
  PublisherDOI

• Data Exploration by Representative Region Selection: Axioms and Convergence

  Mathematics of Operations Research · 2021-03-18 · 1 citations

  article1st authorCorresponding

  We present a new type of unsupervised learning problem in which we find a small set of representative regions that approximates a larger data set. These regions may be presented to a practitioner along with additional information in order to help the practitioner explore the data set. An advantage of this approach is that it does not rely on cluster structure of the data. We formally define this problem, and we present axioms that should be satisfied by functions that measure the quality of representatives. We provide a quality function that satisfies all of these axioms. Using this quality function, we formulate two optimization problems for finding representatives. We provide convergence results for a general class of methods, and we show that these results apply to several specific methods, including methods derived from the solution of the optimization problems formulated in this paper. We provide an example of how representative regions may be used to explore a data set.

  PublisherDOI

• Monge Properties, Optimal Greedy Policies, and Policy Improvement for the Dynamic Stochastic Transportation Problem

  INFORMS journal on computing · 2020-10-13 · 2 citations

  article1st authorCorresponding

  We consider a dynamic, stochastic extension to the transportation problem. For the deterministic problem, there are known necessary and sufficient conditions under which a greedy algorithm achieves the optimal solution. We define a distribution-free type of optimality and provide analogous necessary and sufficient conditions under which a greedy policy achieves this type of optimality in the dynamic, stochastic setting. These results are used to prove that a greedy algorithm is optimal when planning a type of air-traffic management initiative. We also provide weaker conditions under which it is possible to strengthen an existing policy. These results can be applied to the problem of matching passengers with drivers in an on-demand taxi service. They specify conditions under which a passenger and driver should not be left unassigned.

  PublisherDOI

• Quantity-Contingent Auctions and Allocation of Airport Slots

  Transportation Science · 2020 · 29 citations

  • Computer Science
  • Operations research
  • Computer Science

  In this paper, we define and investigate quantity-contingent auctions. Such auctions can be used when there exist multiple units of a single product and the value of a set of units depends on the total quantity sold. For example, a road network or airport will become congested as the number of users increase so that a permit for use becomes more valuable as the total number allocated decreases. A quantity-contingent auction determines both the number of items sold and an allocation of items to bidders. Because such auctions could be used by bidders to gain excessive market power, we impose constraints limiting market power. We focus on auctions that allocate airport arrival and departure slots. We propose a continuous model and an integer programming model for the associated winner determination problem. Using these models, we perform computational experiments that lend insights into the properties of the quantity-contingent auction.

  PublisherDOI

• Facets of the Stochastic Network Flow Problem

  SIAM Journal on Optimization · 2020-01-01 · 3 citations

  article1st authorCorresponding

  We study a type of network flow problem that we call the minimum-cost F-graph flow problem. This problem generalizes the typical minimum-cost network flow problem by allowing the underlying network to be a directed hypergraph rather than a directed graph. This new problem is pertinent because it can be used to model network flow problems that occur in a dynamic, stochastic environment. We formulate this problem as an integer program, and we study specifically the case where every node has at least one outgoing edge with no capacity constraint. We show that even with this restriction, the problem of finding an integral solution is NP-hard. However, we can show that all of the inequality constraints of our formulation are either facet-defining or redundant.

  PublisherDOI

Frequent coauthors

• Michael O. Ball

  15 shared

• David J. Lovell

  5 shared

• Mark Hansen

  2 shared

• Yulin Liu

  2 shared

• Jean‐Philippe P. Richard

  University of Minnesota

  2 shared

• Ankur Mani

  2 shared

Labs

• Smith and Amazon Collaborative ResearchPI

Education

• Ph.D., Applied Mathematics & Statistics, and Scientific Computation

  University of Maryland at College Park

  2018

• Bachelor of Sciences, Mathematics

  University of Nebraska-Lincoln

  2013