A coordinated effort to formalize the mathematics of operations research in Lean 4.
Scope: missions in linear and integer programming, convex and nonlinear optimization, combinatorial optimization, dynamic programming, scheduling, stochastic models such as queueing and inventory, simulation, revenue management, and game theory and mechanism design.
Consumer Privacy in Online Retail Supply Chains I: For a Single Firm the GDPR Raises Profit and Social Welfare, and Raises Consumer Surplus Exactly When v < 2M√(1−δ)/(Z√F)Research Paper
The European Union's General Data Protection Regulation (GDPR, in force since 2018) shifted control over personal data from firms to consumers: data may be collected and processed only with the consumer's consent, and consent may be given for some uses and withheld for others. Online retailers use consumer data in two opposite ways. Saved shopping carts,…
Multi-Product Dynamic Pricing with Reference Effects Under Logit Demand 1: Without Gain-Seeking Products the Myopic Price Path Converges to a Steady State; with One the Myopic Policy Has NoneResearch Paper
Consumers judge a price against a reference price, an internal benchmark formed from the prices they have seen before. A price above the reference is perceived as a loss and depresses demand; a price below it is perceived as a gain and stimulates it. Because today's prices shape tomorrow's reference prices, a seller facing reference effects solves a dynamic…
Telehealth in Acute Care: Pay Parity and Patient Access: Under Pay Parity, for Duplicate-Visit Probability q in (q†, q*), Telehealth Maximizes Provider Revenue but Office-Only Maximizes AccessResearch Paper
During the COVID-19 pandemic most U.S. payers began reimbursing a telehealth consultation at the same rate as the corresponding office visit, a policy called pay parity. Many states have since made it permanent, on the stated ground that it improves patients' access to care. In acute care (primary and urgent care) a telehealth visit sometimes cannot resolve…
Acquihiring for Monopsony Power 3: With a Uniform Private Benefit on [0, B], the Acquirer Acquihires Exactly When v_a > v_s + (7 − 3√5)B/2 and Otherwise Hires DirectlyResearch Paper
Large technology firms often buy small startups mainly to hire their staff, a practice known as acquihiring. Antitrust authorities review acquisitions that may reduce competition in product markets, but an acquihire can instead reduce competition in the labour market: a firm that buys the employer of a scarce worker no longer has to outbid that employer for…
Acquihiring for Monopsony Power 1: In the Baseline Model the Acquirer Hires the Employee Only Through Acquihiring, Exactly When v_a > v_s, and the Employee Prefers Direct HiringResearch Paper
Technology firms often buy small startups mainly to obtain their staff, a practice known as acquihiring; Chatterji and Patro (2014) identify 98 such deals between 2009 and 2013. Bar-Isaac, Johnson and Nocke (CRC TR 224 Discussion Paper No. 500, 2024, www.crctr224.de; published in Management Science, 2025, doi:10.1287/mnsc.2024.04604) study the labour-market…
Capturing the Benefits of Autonomous Vehicles in Ride-Hailing: The Role of Market Configuration 2: On Independent Platforms, AV-HV Coexistence Raises the Price When Labor Is Elastic EnoughResearch Paper
Ride-hailing platforms are expected to add autonomous vehicles (AVs) to fleets of human-driven vehicles (HVs). A common expectation is that AVs, by removing the driver's wage, will make rides cheaper. Lian and van Ryzin (SSRN 3716491; published in Management Science, 2025, DOI 10.1287/mnsc.2020.03112) show that the answer depends on the market configuration.…
Coopetition in Platform-Based Retailing: On the Platform’s Entry II: In the Protective Regime the Seller Underinvests to Deter Entry for Intermediate Product Base ValuesResearch Paper
Online marketplaces such as Amazon are both the storefront of third-party sellers and a retailer in their own right. A platform that observes how well a seller's product sells can later launch a competing product in the same space, and empirical work finds that it does so selectively: Zhu and Liu (2018, Strategic Management Journal) report that Amazon is…
Robustifying Conditional Portfolio Decisions via Optimal Transport 1: With a Singleton Fiber and Only Positive Fiber Mass Required, the Worst-Case Conditional Expected Loss Collapses to sup_y ℓ(y)Research Paper
A portfolio manager who observes side information (a covariateX, such as a market indicator) before choosing a portfolio wants to control the loss of the asset returns Yconditional on the observed covariate value x0. The conditional law of Y given X=x0 is unknown and must be learned from N historical pairs (xi,yi). When x0 itself…
Capturing the Benefits of Autonomous Vehicles in Ride-Hailing: The Role of Market Configuration 3: A Monopoly AV Supplier on a Common Platform Chooses a Newsvendor-Fractile CapacityResearch Paper
Ride-hailing platforms are expected to add autonomous vehicles (AVs) to fleets that today consist of human drivers. AVs have a cost advantage: a human driver must earn a reservation wage for every hour on the platform, while an AV owner pays a fixed capital cost per calendar hour and a smaller variable cost per operating hour. Whether that advantage reaches…
Simultaneous vs. Sequential: Optimal Assortment Recommendation in Multistore Retailing I: Under Simultaneous Recommendation, Offering the Single-Store Optimal Assortment in Every Store Is OptimalResearch Paper
Retail chains increasingly run their physical stores and online channels as one network. A customer who walks into a store and finds nothing she wants can be shown products stocked in other stores of the chain, to be shipped or picked up later. This raises a basic assortment question: when every store can point its customers to the rest of the network,…
Platform Financing vs. Trade Credit for Lending to Third-Party Sellers III: At Low Uncertainty the Platform's Optimal Loan Rate Is Below the Risk-Free Rate Exactly When Production Cost Is HighResearch Paper
Small third-party sellers on e-commerce marketplaces often lack working capital to buy inventory. Two financing channels compete for them. Under trade credit, the supplier ships now and is paid after sales, bearing the seller's default risk. Under platform financing, the marketplace itself (Amazon Lending, Alibaba's microloans, JD's supply-chain finance)…
Common Ownership, Competition, and Corporate Governance: Equilibrium Common Ownership Decreases in Managers' Ability to Divert Profits and Increases in the Cost of MonitoringResearch Paper
Large asset managers now hold significant stakes in many competing firms at once. When the same investor owns shares in rival firms, managers who weigh that investor's interests have an incentive to compete less hard, which raises industry profits at consumers' expense. This "softening of competition" effect is the subject of a large empirical and antitrust…
Selling Data to Marketers 2: In Selling Data Appends with Costs Below the Prior Mean, the High-Cost Plan Is Efficient and the Low-Cost Plan Carries Less Information If φ < 1/2 and More If φ > 1/2Research Paper
Data brokers sell information about consumers to firms that decide whom to target: a lender deciding whom to offer credit, a retailer deciding whom to mail. Two products dominate this market. A marketing list sells contact information together with an analysis, so that a firm cannot reach the consumer at all without buying. A data append adds information…
Third-Degree Price Discrimination in Two-Sided Markets: Seller-Side Discrimination Raises Platform Profit, Buyer and Low-Type Seller Surplus and Total Welfare; High-Type Sellers Gain iff b > b̂Research Paper
Third-degree price discrimination means charging different prices to identifiable groups of customers. Its welfare effects are a classical question in industrial organization. With linear demands and every market served under uniform pricing, discrimination leaves total output unchanged and lowers total welfare (Pigou 1920). More generally, welfare can rise…
Reducing Marketplace Interference Bias Via Shadow Prices 2: In Expectation the Shadow-Price Estimator Equals the Derivative of the Finite-Sample Partial-Treatment Value FunctionResearch Paper
Two-sided platforms (ride hailing, delivery, freight, cloud capacity) test product changes with user-split experiments: each arriving demand unit is assigned to treatment with probability ρ and to control otherwise. Treated and control units then compete for the same supply, so the outcome of one group depends on the size of the other. This interference b…
Revenue Sharing at Music Streaming Platforms II: The Core of the Streaming Game Is the Set of Allocations Splitting Each User's Payment Among the Artists That User PlayedResearch Paper
Music streaming platforms collect a subscription fee from each user and have to pass the revenue on to the artists whose content was played. The two rules used in practice are pro-rata (each artist is paid in proportion to their total number of streams, over all users) and user-centric (each user's fee is split among the artists that user played, in…
Global Sensitivity Analysis via Optimal Transport 1: An OT-Based Sensitivity Measure Is Maximal Exactly When the Output Is a Function of the InputResearch Paper
Global sensitivity analysis asks how much of the uncertainty in the output Y of a model, a simulator or a learned predictor is attributable to an input X. A standard way to answer is to compare the law of Y with its law after X is revealed, and to average that discrepancy over X. Variance-based indices (Sobol' 1993; Wagner 1995) compare conditional and…
Coopetition in Platform-Based Retailing: On the Platform’s Entry I: The Seller Prefers Entry Exactly Above Thresholds in Spillover and Referral Rate and Outside a Band of Competition IntensityResearch Paper
Large online marketplaces such as Amazon, JD.com and Walmart.com act both as a sales channel for third-party sellers and as a retailer of their own. A platform watches which third-party products sell well and may start selling the same product itself, a move called platform entry. Over ten months of 2013–2014, Amazon entered about 3% of the products first…
Selling Data to Marketers 1: In Selling Marketing Lists, the Low-Cost Plan Is Efficient and the High-Cost Plan Carries More Information Than the Benchmark If φ < 1/2 and Less If φ > 1/2Research Paper
Data brokers such as Acxiom, Nielsen and Oracle, and data generators such as Google and Facebook, sell processed consumer data to firms that use it to decide whom to target, whom to lend to, or which market to enter. The U.S. Federal Trade Commission's 2014 report on data brokers distinguishes two products: a marketing list, which carries contact…
Round-Robin Beyond Additive Agents: Existence and Fairness of Approximate Equilibria 2: For Submodular Valuations the Bluff Profile Is a Tight 1/2-Approximate Pure Nash Equilibrium of Round-RobinResearch Paper
Dividing indivisible goods among agents without money is a basic problem of fair division. The most widely used procedure is Round-Robin: the agents take turns, in a fixed order, picking their favourite remaining good. For agents with additive valuations, Round-Robin run on the agents' true preferences produces an allocation that is envy-free up to one good…
The Value of Information Design in Supply Chain Management III: With Overlapping Supports, the Optimal Policy Goes from Partial to No to Full Disclosure as the Critical Ratio RisesResearch Paper
A retailer often knows more about her own demand than the supplier who must produce for her. In a decentralized supply chain the supplier sets a production quantity before demand is realized, and an under-informed supplier produces too little or too much. The retailer can shape what the supplier believes by committing in advance to a disclosure policy, a…
Revenue Sharing at Music Streaming Platforms I: Homogeneity and Additivity Characterize Weighted Indices; Adding Equal Individual or Global Impact of Users Gives Pro-Rata or User-CentricResearch Paper
Music streaming platforms collect a subscription fee from each user and pay part of the revenue to the artists whose content was played. The industry standard is the pro-rata method: the revenue pool is divided in proportion to each artist's total number of streams across all users. A competing method, the user-centric method, divides each user's own…
Multitask Learning and Bandits via Robust Statistics: High-Probability ℓ1 Error Bound for the Robust Multitask (Trimmed Mean + LASSO) Estimator under Sparse HeterogeneityResearch Paper
Service providers such as hospitals or retail stores each fit a predictive model on their own data, and many of them have few observations. The models of different providers are usually similar but not identical: most coefficients agree, while each provider deviates from the common model in a few features of its own. Multitask learning asks how data from…
Retention or Acquisition? Behavior-Based Quality Disclosure I: Retention-Focused Disclosure Thresholds and Higher Second-Period Profits under Behavior-Based DisclosureResearch Paper
Firms that hold purchase-history data can treat past customers and a rival's customers differently. The economics of behavior-based pricing (BBP), charging different prices to the two groups, is well developed: in a symmetric duopoly, poaching the rival's customers with discounts intensifies competition and usually lowers profits (Fudenberg and Tirole 2000; …
Thompson Sampling with Information Relaxation Penalties II: The IRS Performance Bounds Are Monotone, W^TS ≥ W^IRS.FH ≥ W^IRS.V-Zero and W^TS ≥ W^IRS.V-EMaxResearch Paper
In a Bayesian multi-armed bandit a decision maker pulls one of K arms in each of T periods, observes the reward, and updates a prior belief about each arm's unknown parameter. The best achievable expected total reward, V∗(T,y), is the value of a dynamic program over beliefs, and for all but the smallest instances it cannot be computed. Upper bounds on V∗…