SPSaša Pekeč

Saša Pekeč

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Saša Pekeč has a total of 21 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Beyond Truthful Reporting: Robust Strategies for Worst-Case Payoff Maximization in Large Markets

    by Saša Pekeč, Chenxi Xu

    In many large-scale markets mediated by digital platforms, such as advertising, energy trading, transportation logistics, and spectrum auctions, platforms suggest that participants report true valuations to simplify strategic complexity. This recommendation is widely followed by smaller bidders who lack resources for sophisticated strategic analysis. This paper shows that simple deviations from truthful reporting can improve outcomes for such participants. Using a robust optimization framework, we derive practical bidding strategies for generalized first-price, generalized second-price, and core-selecting combinatorial auctions. Our distribution-free approach maximizes worst-case payoffs over a set of plausible competitor bids, without requiring detailed information about rival behavior. We demonstrate that simple strategies, specifically shading bids on preferred bundles, consistently outperform truthful bidding when bidders have straightforward valuation structures. This insight persists even in complex settings where closed-form optimal policies cannot be derived. The results provide actionable guidance for market participants with limited information who lack resources for deploying sophisticated bidding algorithms, while offering market designers insight into how participants might deviate from recommended strategies.

  • Open Access

    Clearing Multi-Unit Markets with Simple Pricing

    by Ali Makhdoumi, Saša Pekeč, Xinchang Wang

    We study multi-unit markets where a seller allocates multiple units of an item to buyers with different valuation functions via simple pricing. When the valuation functions are submodular, a Walrasian equilibrium with a single (per-unit) price exists, i.e., each buyer obtains an allocation that maximizes her utility and the market clears. However, market clearing may fail, and a Walrasian equilibrium might not exist even under slight deviations from submodularity. To study how such equilibria can be approximated for more general classes of valuations, we introduce the τ-approximate Walrasian equilibrium: a market-clearing allocation and a single price for which each buyer's utility is not necessarily maximum, but it is weakly larger than the maximum utility under τ-scaled valuations. We establish that any τ-approximate Walrasian equilibrium guarantees at least a τ-fraction of the optimal social welfare, which motivates our main question: what is the tightest approximation of the Walrasian equilibrium achievable with a single price? We provide a Pseudo-Polynomial Time Approximation Scheme (PPTAS) that finds the tightest approximation for general valuation functions. Furthermore, for single-peaked marginal valuation functions, which generalizes both submodular and supermodular valuation functions, we utilize the demand structure to design Fully Polynomial-Time Approximation Scheme (FPTAS). We go beyond a single price and, interestingly, establish a structural separation: while a single anonymous price suffices for single-minded buyers, non-anonymous pricing (specifically, offering distinct per-unit prices to two buyer groups) is necessary to find the tightest relaxation for buyers with general single-peaked marginal valuations.

  • Open Access

    Ballot Design and Electoral Outcomes: The Role of Candidate Order and Party Affiliation

    by Alessandro Arlotto, Alexandre Belloni, Fei Fang, Saša Pekeč

    We use causal inference to study how designing ballots with and without party designations impacts electoral outcomes when partisan voters rely on party-order cues to infer candidate affiliation in races without designations. If the party orders of candidates in races with and without party designations differ, these voters might cast their votes incorrectly. We identify a quasi-randomized natural experiment with contest-level treatment assignment pertaining to North Carolina judicial elections and use double machine learning to accurately capture the magnitude of such incorrectly cast votes. Using precinct-level election and demographic data, we estimate that 11.8% (95% confidence interval: [4.0%, 19.6%]) of democratic partisan voters and 15.4% (95% confidence interval: [7.8%, 23.1%]) of republican partisan voters cast their votes incorrectly due to the difference in party orders. Our results indicate that ballots mixing contests with and without party designations mislead many voters, leading to outcomes that do not reflect true voter preferences. To accurately capture voter intent, such ballot designs should be avoided.

  • Approximate Submodularity in Network Design Problems

    by Levi DeValve, Saša Pekeč, Yehua Wei

    The paper “Approximate Submodularity in Network Design Problems” studies general network design problems where a firm strategically selects its network to better match supply and demand in the future. The paper observes that the arcs in network design problems are cover modular, i.e., approximately substitutes with each other, in the sense that local changes in the objective function can be used to bound global changes. The cover modularity is then applied to prove that a set of simple and intuitive heuristics achieve constant factor approximation guarantees in network design problems. Furthermore, the paper demonstrates cover modularity is also present in a general class of linear programming formulations.


  • Courteous or Crude? Managing User Conduct to Improve On-Demand Service Platform Performance

    by Yunke Mai, Bin Hu, Saša Pekeč

    In this paper, we study how an on-demand service platform could improve its performance through managing user conduct. In such a platform, service providers may reject certain platform-proposed service requests, and their responses, in turn, incentivize users to adjust their conduct. We develop an evolutionary game theory model of user conduct and provider responses that shows that the platform could improve user conduct through either setting a low wage for service providers or implementing priority matching. Building upon these results, we further model providers and users joining and leaving the platform by once again utilizing the evolutionary game theory approach. We find that wage setting alone is a blunt instrument to improve platform performance via managing user conduct, whereas supplementing the wage decision with priority matching could overcome its limitations and serve as an effective strategy to further improve platform performance in terms of growth and profitability. This finding suggests that matching prioritization could be an important strategy for managing platforms with user and provider heterogeneities. In addition, our analysis and results also demonstrate the potential of the evolutionary game theory approach for analyzing the impact of pricing and matching decisions on the performance of large markets. This paper was accepted by Gabriel Weintraub, revenue management and market analytics. Supplemental Material: The online appendix is available at https://doi.org/10.1287/mnsc.2022.4391 .


  • Top Cited Papers

    Revenue-Sharing Allocation Strategies for Two-Sided Media Platforms: Pro-Rata vs. User-Centric

    by Saeed Alaei, Ali Makhdoumi, Azarakhsh Malekian, Saša Pekeč

    We consider a two-sided streaming service platform that generates revenues by charging users a subscription fee for unlimited access to the content and compensates content providers (artists) through a revenue-sharing allocation rule. Platform users are heterogeneous in both their overall consumption and the distribution of their consumption over different artists. We study two primary revenue allocation rules used by market-leading music streaming platforms—pro-rata and user-centric. With pro-rata, artists are paid proportionally to their share of the overall streaming volume, whereas with user-centric, each user’s subscription fee is divided proportionally among artists based on the consumption of that user. We characterize when these two allocation rules can sustain a set of artists on the platform and compare them from both the platform’s and the artists’ perspectives. In particular, we show that, despite the cross-subsidization between low- and high-streaming-volume users, the pro-rata rule can be preferred by both the platform and the artists. Furthermore, the platform’s problem of selecting an optimal portfolio of artists is NP-complete. However, by establishing connections to the knapsack problem, we develop a polynomial time approximation scheme (PTAS) for the optimal platform’s profit. In addition to determining the platform’s optimal revenue allocation rule in the class of pro-rata and user-centric rules, we consider the optimal revenue allocation rule in the class of arbitrary rules. Building on duality theory, we develop a polynomial time algorithm that outputs a set of artists so that the platform’s profit is within a single artist’s revenue from the optimal profit. This paper was accepted by Gabriel Weintraub, revenue management and market analytics. Supplemental Material: The online appendices are available at https://doi.org/10.1287/mnsc.2022.4307 .


  • Optimal Price/Advertising Menus for Two-Sided Media Platforms

    by Levi DeValve, Saša Pekeč

    Ad-supported media platforms often engage price-sensitive consumers by offering content for free while reaping revenue from advertisers. Many consumers are also sensitive to advertising, however, so platforms need to carefully balance their advertising quantities with the prices they offer to consumers. Several platforms use menu pricing to achieve this goal: they offer consumers a choice between a free-use/ad-supported option or a paid-use/no-ad option. In “Optimal Price/Advertising Menus for Two-Sided Media Platforms,” DeValve and Pekeč use Lagrangian duality to establish the optimality of such menus in a model of consumers with heterogeneous advertising sensitivity, offering both theoretical justification for their use in practice as well as guidance on when and how to set positive prices. Moreover, the optimal menu characterization implies that platforms may have more of an incentive to offer menu pricing under competition (further explaining their prevalence in the competitive digital media market) and platforms primarily cater to consumers with low advertising sensitivity (suggestive of the proliferation of online advertising observed in practice).


  • A Primal-Dual Approach to Analyzing ATO Systems

    by Levi DeValve, Saša Pekeč, Yehua Wei

    We study assemble-to-order (ATO) problems from the literature. ATO problems with general structure and integrality constraints are well known to be difficult to solve, and we provide new insight into these issues by establishing worst-case approximation guarantees through primal-dual analyses and linear programming (LP) rounding. First, we relax the one-period ATO problem using a natural newsvendor decomposition and use the dual solution for the relaxation to derive a lower bound on optimal cost, providing a tight approximation guarantee that grows with the maximum product size in the system. Then, we present an LP rounding algorithm that achieves both asymptotic optimality as demand grows large, and a 1.8 approximation factor for any problem instance. In addition to theoretical guarantees, we perform comprehensive numerical simulations and find that our rounding algorithm outperforms existing techniques and is close to optimal. Finally, we demonstrate that our one-period LP rounding results can be used to develop an asymptotically optimal integral policy for dynamic ATO problems with backlogging and identical component lead-times. This paper was accepted by Yinyu Ye, optimization.


  • Efficient Allocation and Pricing of Multifeatured Items

    by Ozan Candogan, Saša Pekeč

    We introduce feature valuations, a new class of valuations that compactly capture preferences of agents who value items based on the features they possess. Such preferences are relevant in many important practical settings, such as Internet advertising markets (where impressions have associated attributes), labor markets (where workers’ skills/titles correspond to features), and long-term bond markets (where coupon payments can be associated with features). We focus on settings where features can be organized in a tree network, and features possessed by items are captured by paths on the tree. In such settings, under appropriate consistency conditions, we establish that Walrasian equilibria exist. In addition, we provide a computationally tractable price update algorithm that terminates with market-clearing prices that support an efficient outcome. The e-companion is available at https://doi.org/10.1287/mnsc.2017.2917 . This paper was accepted by Yinyu Ye, optimization.