INFORMS Workshop on Market Design

About the Workshop

The 2026 INFORMS Workshop on Market Design will be held on July 6th, 2026 in conjunction with the EC 2026 conference. The workshop will bring together researchers and practitioners that work on market design, and will represent a broad range of perspectives from more theoretical to more applied and empirical work. As with previous iterations, the workshop is sponsored and organized by the INFORMS Section on Auctions and Market Design, i.e., the workshop is a successor of the earlier editions, going all the way to 2018.

Workshop Theme

Market design is a field of applied and theoretical research that sits comfortably on the intersection of computer science, economics, and operations research. In recent decades, the theory and applications of market design have blossomed. In this workshop, we will focus on a set of promising, new applications of market design. In particular, we are interested in applications of market design which involve optimization with complex allocation constraints, vast datasets and machine learning, and dynamic pricing issues. We also want to explore research areas which, despite receiving theoretical attention, have not yet made an impact on practice. We are also interested in empirical work on evaluating mechanisms and approaches in practice. Topics include but are not limited to:

• Machine learning and generative AI in market design

• Mathematical optimization in markets

• Pricing and competitive equilibria in markets

• Iterative multi-object auctions

• Matching with constraints and complex preferences

Schedule

9:00 - 9:45 — Sasa Pekec: Market Clearing Beyond Submodularity

• Abstract: The ability to coordinate market sides at speed and scale is a cornerstone of modern platform business models and resource allocation systems more broadly. At its core, market clearing involves two fundamental tasks: allocation and pricing. This talk develops primal-dual based approaches to tackle each task in settings that go beyond the submodular structure typically required for tractable, optimal solutions. On the allocation side, we study multi-sided matching under supply and demand uncertainty with capacity constraints. The key technical contribution is identifying a relaxation of submodularity that emerges naturally from primal-dual analysis of a linear programming formulation of the allocation task. Leveraging this, we establish constant-factor approximation guarantees for a broad class of problems that subsume not only matching but also flexibility design, facility location, and subset selection. On the pricing side, we consider when anonymous per-unit prices can support an approximately efficient allocation. Motivated by settings such as standardized procurement, cloud resource pricing, and posted-price sales of homogeneous goods, we leverage a linear programming representation of the market-clearing objective and again utilize the primal-dual approach to define a suitable approximation of Walrasian equilibrium. Building on this, we design and analyze algorithms that yield the uniform price that clears the market and supports an approximate equilibrium with corresponding welfare guarantees.

9:45 - 10:30 — Xizhi Tan: Forging Self-Funded Marketplaces among Strategic Agents

• Abstract: Forging Self-Funded Marketplaces among Strategic Agents  Abstract: Many modern platforms and decentralized systems operate as self-funded ecosystems where the revenue generated by participants must cover their procurement costs. This introduces a complex autarkic mechanism design problem with an endogenous budget constraint, which fundamentally distinguishes it from classic budget-feasible procurement. We show that when agents are strategic and hold private costs, demanding exact budget balance leads to severe impossibility results, preventing truthful mechanisms from achieving any bounded approximation to the first-best benchmark.  To overcome these fundamental barriers, we explore two complementary theoretical frameworks. First we show that by relaxing the truthfulness constraint and employing a sequential Best and Final Offer (BAFO) protocol, we can match theoretical limits, achieving an $O(\log \mathcal{V})$ approximation to the optimal value in every subgame perfect equilibrium. Second, to better capture market competitiveness beyond quasi-monopolistic settings, we introduce the Maximin Share (MMS) benchmark and provide a sequential mechanism that achieves a constant-factor approximation.  Finally, applying a resource-augmentation perspective, we demonstrate that allowing a small multiplicative budget relaxation ($\beta > 1$) enables simple, prior-free randomized posted-price mechanisms to achieve constant approximations.

Joint work with Yuan Deng, Vasilis Gkatzelis, Amin Saberi, Grigoris Velegkas, Ellen Vitercik, Song Zuo.

10:30 - 11:00 — Coffee Break

11:00 - 11:45 — Thanh Nguyen: Market Design After Deferred Acceptance

• Abstract: Many central ideas in market design are equilibrium ideas: stability, competitive equilibrium, and the core. Yet the outcomes we implement are discrete. A student attends one school, a worker takes one job, and a committee has actual members rather than fractional shares.  Deferred acceptance resolves this tension beautifully in settings with the right preference structure: a discrete algorithm implements a compelling equilibrium notion. But many important environments fall outside this framework. Preferences may be ordinal but not substitutable, constraints may be combinatorial, outcomes may be public, and exact equilibrium may fail to exist.  This talk discusses a broader approach to market design. Instead of starting with an algorithm, I start with an equilibrium or fixed-point object that captures the relevant economic forces, and then convert it into a discrete outcome while preserving guarantees such as stability, fairness, or representation. I will describe techniques for defining equilibrium with ordinal preferences and for rounding equilibrium objects without destroying their economic content. Applications include fair allocation, stable matching, and collective social choice.

11:45 - 12:30 — Lawrence M.  Ausubel: Auction Design for Artificial-Intelligence-Based Sponsored Search

• Abstract: Sponsored search and keyword auctions—by which advertisers place bids on keywords and winning bidders are awarded sponsored links at the top of search pages—have been enduring features of internet search for the past quarter century. However, emerging artificial intelligence technologies such as ChatGPT pose significant challenges to the existing model for monetizing internet search, as there is a growing mismatch. Specifically, the current “sponsored” search engine output consists of an ordered list of advertisers’ URLs at the top of the search page, whereas the “organic” search engine output is evolving from ordered lists of links to paragraphs of free-form text directly addressing the consumer’s query. As a result, it is plausible that consumers will largely disregard the sponsored links. This paper explores a possible transformation to the auction design of sponsored internet search that may ensue. Specifically, an advertiser would bid to “influence” the output of the search provider in a direction favorable to the advertiser—for example, a higher bid could correspond to a more positive description of the advertiser’s product in the AI-written text response. We conceptualize the mechanism design problem of a search provider using a trading model among stakeholders. The search provider allows stakeholders to “buy” increased purchase probability by paying into the trading mechanism or to “sell” by ceding purchase probability to other stakeholders. The mechanism designer maximizes revenues subject to a constraint that the increases and decreases in purchase probability can be no greater than specified amounts, capturing that there is a limit on how much the assessments of stakeholders can be allowed to change while maintaining the search provider’s credibility with consumers. In the solution of one formulation, purchase probability is shifted entirely toward the single stakeholder bidding the most and is shifted entirely away from the stakeholder bidding the least.

12:30-2:00 —  Lunch and poster session

Speakers

Venue

Auditorium Antonianum

Viale Manzoni, 1
00185, Rome
Tel/Fax +39.06.45582593
Website: www.auditoriumantonianum.it

Organizers

Contact: informs-market-design-workshop@googlegroups.com