Marco Scarsini
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Marco Scarsini has a total of 115 co-authors. Here are the top co-authors by number of shared publications:
- Alfred Müller
Department Mathematik, Universität Siegen, 57072 Siegen, Germany
13 shared publications - Roberto Cominetti
Facultad de Ingeniería y Ciencias, Universidad Adolfo Ibáñez, Santiago 7941169, Chile
12 shared publications Moshe Shaked
9 shared publicationsBruno Bassan
University of Rome 1
7 shared publications- Tristan Tomala
HEC Paris and GREGHEC 1 rue de la Libération 78351 Jouy-en-Josas France
7 shared publications - Marc Schröder
Maastricht University
6 shared publications
Latest Publications
- 1 August 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Mathematics of Operations Research
- Vol. 50 (3) , 2157-2174
Correlated Equilibria in Large Anonymous Bayesian Games
We consider multipopulation Bayesian games with a large number of players. Each player aims at minimizing a cost function that depends on this player’s own action, the distribution of players’ actions in all populations, and an unknown state parameter. We study the nonatomic limit versions of these games and introduce the concept of Bayes correlated Wardrop equilibrium, which extends the concept of Bayes correlated equilibrium to nonatomic games. We prove that Bayes correlated Wardrop equilibria are limits of action flows induced by Bayes correlated equilibria of the game with a large finite set of small players. For nonatomic games with complete information admitting a convex potential, we prove that the set of correlated and of coarse correlated Wardrop equilibria coincide with the set of probability distributions over Wardrop equilibria and that all equilibrium outcomes have the same costs. We get the following consequences. First, all flow distributions of (coarse) correlated equilibria in convex potential games with finitely many players converge to mixtures of Wardrop equilibria when the weight of each player tends to zero. Second, for any sequence of flows satisfying a no-regret property, its empirical distribution converges to the set of distributions over Wardrop equilibria, and the average cost converges to the unique Wardrop cost. Funding: This work was partially supported by European Cooperation in Science and Technology Action 16228 GAMENET. F. Koessler acknowledges the support of the Agence Nationale de la Recherche [Grant StratCom ANR-19-CE26-0010-01]. M. Scarsini acknowledges the support of the Gruppo Nazionale per l’Analisi Matematica, la Probabilità e le loro Applicazioni project [Grant CUP_E53C22001930001], the Ministero dell’Università e della Ricerca Progetti di Rilevante Interesse Nazionale [Grant 2022EKNE5K], and the European Union-Next Generation EU, component M4C2, investment 1.1 (Ministero dell’Università e della Ricerca Progetti di Rilevante Interesse Nazionale Piano Nazionale di Ripresa e Resilienza) [Grant P2022XT8C8]. T. Tomala gratefully acknowledges the support of the HEC foundation and Agence Nationale de la Recherche/Investissements d’Avenir [Grant ANR-11-IDEX-0003/Labex Ecodec/ANR-11-LABX-0047].
- 1 August 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Mathematics of Operations Research
- Vol. 50 (3) , 2157-2174
Correlated Equilibria in Large Anonymous Bayesian Games
We consider multipopulation Bayesian games with a large number of players. Each player aims at minimizing a cost function that depends on this player’s own action, the distribution of players’ actions in all populations, and an unknown state parameter. We study the nonatomic limit versions of these games and introduce the concept of Bayes correlated Wardrop equilibrium, which extends the concept of Bayes correlated equilibrium to nonatomic games. We prove that Bayes correlated Wardrop equilibria are limits of action flows induced by Bayes correlated equilibria of the game with a large finite set of small players. For nonatomic games with complete information admitting a convex potential, we prove that the set of correlated and of coarse correlated Wardrop equilibria coincide with the set of probability distributions over Wardrop equilibria and that all equilibrium outcomes have the same costs. We get the following consequences. First, all flow distributions of (coarse) correlated equilibria in convex potential games with finitely many players converge to mixtures of Wardrop equilibria when the weight of each player tends to zero. Second, for any sequence of flows satisfying a no-regret property, its empirical distribution converges to the set of distributions over Wardrop equilibria, and the average cost converges to the unique Wardrop cost. Funding: This work was partially supported by European Cooperation in Science and Technology Action 16228 GAMENET. F. Koessler acknowledges the support of the Agence Nationale de la Recherche [Grant StratCom ANR-19-CE26-0010-01]. M. Scarsini acknowledges the support of the Gruppo Nazionale per l’Analisi Matematica, la Probabilità e le loro Applicazioni project [Grant CUP_E53C22001930001], the Ministero dell’Università e della Ricerca Progetti di Rilevante Interesse Nazionale [Grant 2022EKNE5K], and the European Union-Next Generation EU, component M4C2, investment 1.1 (Ministero dell’Università e della Ricerca Progetti di Rilevante Interesse Nazionale Piano Nazionale di Ripresa e Resilienza) [Grant P2022XT8C8]. T. Tomala gratefully acknowledges the support of the HEC foundation and Agence Nationale de la Recherche/Investissements d’Avenir [Grant ANR-11-IDEX-0003/Labex Ecodec/ANR-11-LABX-0047].
- 1 March 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 73 (2) , 672-688
Ordinary and Prophet Planning Under Uncertainty in Bernoulli Congestion Games
Uncertain Demand in Transportation Reduces Worst-Case Congestion In “Ordinary and Prophet Planning Under Uncertainty in Bernoulli Congestion Games,” Cominetti, Scarsini, Schröder, and Stier-Moses characterize the inefficiency introduced by self-minded travelers in congested networks, such as those in big cities. When the likelihood that travelers are present in a congested network is smaller, the uncertainty affects the resulting conditions but also what a social planner would do. Contrary to what might be expected, the consequence is a reduction of the worst-case expressions for the inefficiency of the prevailing traffic conditions with respect to the routes of a social planner. This reduction holds under two possible informational assumptions for the social planner. An ordinary planner only has statistical information based on past data in terms of the expected number of travelers, whereas a prophet planner is informed of the participation outcomes right when travelers start their trips.
- 1 March 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 73 (2) , 672-688
Ordinary and Prophet Planning Under Uncertainty in Bernoulli Congestion Games
Uncertain Demand in Transportation Reduces Worst-Case Congestion In “Ordinary and Prophet Planning Under Uncertainty in Bernoulli Congestion Games,” Cominetti, Scarsini, Schröder, and Stier-Moses characterize the inefficiency introduced by self-minded travelers in congested networks, such as those in big cities. When the likelihood that travelers are present in a congested network is smaller, the uncertainty affects the resulting conditions but also what a social planner would do. Contrary to what might be expected, the consequence is a reduction of the worst-case expressions for the inefficiency of the prevailing traffic conditions with respect to the routes of a social planner. This reduction holds under two possible informational assumptions for the social planner. An ordinary planner only has statistical information based on past data in terms of the expected number of travelers, whereas a prophet planner is informed of the participation outcomes right when travelers start their trips.
- 1 March 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 73 (2) , 879-893
Multivariate Almost Stochastic Dominance: Transfer Characterizations and Sufficient Conditions Under Dependence Uncertainty
Decision with Several Objectives Under Uncertainty Important decisions typically involve multiple objectives and uncertainty. Assessing both multivariate utility and multivariate distributions for the attributes can be challenging. Moreover, big decisions are usually made by boards or committees with members holding divergent views and preferences and facing pressures from different stakeholders. Thus, a full-blown traditional decision analysis that leads to the computation of expected utility is very difficult at best and often not possible. In “Multivariate Almost Stochastic Dominance: Transfer Characterizations and Sufficient Conditions Under Dependence Uncertainty” Müller, Scarsini, Tsetlin, and Winkler develop sufficient conditions for multivariate almost stochastic dominance based on marginal distributions of the attributes or just on their means and variances. Such tools, consistent with normative decision analysis, are useful when making important decisions in today’s fast-moving and often complex world.
- 1 March 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 73 (2) , 879-893
Multivariate Almost Stochastic Dominance: Transfer Characterizations and Sufficient Conditions Under Dependence Uncertainty
Decision with Several Objectives Under Uncertainty Important decisions typically involve multiple objectives and uncertainty. Assessing both multivariate utility and multivariate distributions for the attributes can be challenging. Moreover, big decisions are usually made by boards or committees with members holding divergent views and preferences and facing pressures from different stakeholders. Thus, a full-blown traditional decision analysis that leads to the computation of expected utility is very difficult at best and often not possible. In “Multivariate Almost Stochastic Dominance: Transfer Characterizations and Sufficient Conditions Under Dependence Uncertainty” Müller, Scarsini, Tsetlin, and Winkler develop sufficient conditions for multivariate almost stochastic dominance based on marginal distributions of the attributes or just on their means and variances. Such tools, consistent with normative decision analysis, are useful when making important decisions in today’s fast-moving and often complex world.
- 10 August 2024
- Published by Springer Nature
- p. 309-316
Conditions for the Multivariate Stochastic Order Under Dependence Uncertainty
- 10 August 2024
- Published by Springer Nature
- p. 309-316
Conditions for the Multivariate Stochastic Order Under Dependence Uncertainty
- Open Access
- 8 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 580
Optimal Queueing Regimes
It is well known that in an M/M/1 queueing model where customers strategically decide whether or not to enter a queue, and if and when to renege, under a first-come-first-served regime, customers' selfish behavior produces an outcome that is socially suboptimal. Optimality could be achieved by adopting a different queuing regime. In particular, optimality is achieved by any regime where a new arriving customer is put in a position that is not the last. The priority slots regime, is also universally optimal. Under this regime, there is a sequence of slots indexed by the positive integers. A new arriving customer who enters the queue occupies the available slot with the smallest index. At any given time the customer occupying the slot with the smallest index is being served. A new customer who finds the first slot vacant preempts the current service. Once served, a customer leaves the system and frees the slot she occupied. Authors' addresses: Marco Scarsini, [email protected], Luiss University, Rome, RM, Italy; Eran Shmaya, [email protected], Stony Brook University, Stony Brook, NY, USA.
- 8 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 580
Open AccessOptimal Queueing Regimes
It is well known that in an M/M/1 queueing model where customers strategically decide whether or not to enter a queue, and if and when to renege, under a first-come-first-served regime, customers' selfish behavior produces an outcome that is socially suboptimal. Optimality could be achieved by adopting a different queuing regime. In particular, optimality is achieved by any regime where a new arriving customer is put in a position that is not the last. The priority slots regime, is also universally optimal. Under this regime, there is a sequence of slots indexed by the positive integers. A new arriving customer who enters the queue occupies the available slot with the smallest index. At any given time the customer occupying the slot with the smallest index is being served. A new customer who finds the first slot vacant preempts the current service. Once served, a customer leaves the system and frees the slot she occupied. Authors' addresses: Marco Scarsini, [email protected], Luiss University, Rome, RM, Italy; Eran Shmaya, [email protected], Stony Brook University, Stony Brook, NY, USA.
- Open Access
- 25 March 2024
- Published by Elsevier in Games and Economic Behavior
- Vol. 145, 239-262
Best-response dynamics in two-person random games with correlated payoffs
- 25 March 2024
- Published by Elsevier in Games and Economic Behavior
- Vol. 145, 239-262
Open AccessBest-response dynamics in two-person random games with correlated payoffs
- Open Access
Ordinary and prophet planning under uncertainty in Bernoulli congestion games
We consider an atomic congestion game in which each player $i$ either participates in the game with an exogenous and known probability $p_{i}\in(0,1]$, independently of everybody else, or stays out and incurs no cost. We compute the parameterized price of anarchy to characterize the impact of demand uncertainty on the efficiency of selfish behavior, considering two different notions of a social planner. A prophet planner knows the realization of the random participation in the game; the ordinary planner does not. As a consequence, a prophet planner can compute an adaptive social optimum that selects different solutions depending on the players that turn out to be active, whereas an ordinary planner faces the same uncertainty as the players and can only compute social optima with respect to the player participation distribution. For both planners, we derive the precise price of anarchy, which arises from an optimization problem parameterized by the maximum participation probability $q=\max_{i} p_{i}$. In the case of affine costs, we provide an analytic expression for the ordinary and prophet price of anarchy, parameterized as a function of $q$.Open AccessOrdinary and prophet planning under uncertainty in Bernoulli congestion games
We consider an atomic congestion game in which each player $i$ either participates in the game with an exogenous and known probability $p_{i}\in(0,1]$, independently of everybody else, or stays out and incurs no cost. We compute the parameterized price of anarchy to characterize the impact of demand uncertainty on the efficiency of selfish behavior, considering two different notions of a social planner. A prophet planner knows the realization of the random participation in the game; the ordinary planner does not. As a consequence, a prophet planner can compute an adaptive social optimum that selects different solutions depending on the players that turn out to be active, whereas an ordinary planner faces the same uncertainty as the players and can only compute social optima with respect to the player participation distribution. For both planners, we derive the precise price of anarchy, which arises from an optimization problem parameterized by the maximum participation probability $q=\max_{i} p_{i}$. In the case of affine costs, we provide an analytic expression for the ordinary and prophet price of anarchy, parameterized as a function of $q$. - Open Access
- 2 February 2024
- Published by Elsevier in European Journal of Operational Research
- Vol. 316 (2) , 754-766
Monotonicity of equilibria in nonatomic congestion games
- 2 February 2024
- Published by Elsevier in European Journal of Operational Research
- Vol. 316 (2) , 754-766
Open AccessMonotonicity of equilibria in nonatomic congestion games
- Open Access
Correlated Equilibria in Large Anonymous Bayesian Games
We consider multi-population Bayesian games with a large number of players. Each player aims at minimizing a cost function that depends on this player's own action, the distribution of players' actions in all populations, and an unknown state parameter. We study the nonatomic limit versions of these games and introduce the concept of Bayes correlated Wardrop equilibrium, which extends the concept of Bayes correlated equilibrium to nonatomic games. We prove that Bayes correlated Wardrop equilibria are limits of action flows induced by Bayes correlated equilibria of the game with a large finite set of small players. For nonatomic games with complete information admitting a convex potential, we prove that the set of correlated and of coarse correlated Wardrop equilibria coincide with the set of probability distributions over Wardrop equilibria, and that all equilibrium outcomes have the same costs. We get the following consequences. First, all flow distributions of (coarse) correlated equilibria in convex potential games with finitely many players converge to Wardrop equilibria when the weight of each player tends to zero. Second, for any sequence of flows satisfying a no-regret property, its empirical distribution converges to the set of distributions over Wardrop equilibria and the average cost converges to the unique Wardrop cost.
Open AccessCorrelated Equilibria in Large Anonymous Bayesian Games
We consider multi-population Bayesian games with a large number of players. Each player aims at minimizing a cost function that depends on this player's own action, the distribution of players' actions in all populations, and an unknown state parameter. We study the nonatomic limit versions of these games and introduce the concept of Bayes correlated Wardrop equilibrium, which extends the concept of Bayes correlated equilibrium to nonatomic games. We prove that Bayes correlated Wardrop equilibria are limits of action flows induced by Bayes correlated equilibria of the game with a large finite set of small players. For nonatomic games with complete information admitting a convex potential, we prove that the set of correlated and of coarse correlated Wardrop equilibria coincide with the set of probability distributions over Wardrop equilibria, and that all equilibrium outcomes have the same costs. We get the following consequences. First, all flow distributions of (coarse) correlated equilibria in convex potential games with finitely many players converge to Wardrop equilibria when the weight of each player tends to zero. Second, for any sequence of flows satisfying a no-regret property, its empirical distribution converges to the set of distributions over Wardrop equilibria and the average cost converges to the unique Wardrop cost.
- 1 July 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (4) , 1136-1153
Product Ranking in the Presence of Social Learning
This paper studies product ranking mechanisms of a monopolistic online platform in the presence of social learning. The products’ quality is initially unknown, but consumers can sequentially learn it as online reviews accumulate. A salient aspect of our problem is that consumers, who want to purchase a product from a list of items displayed by the platform, incur a search cost while scrolling down the list. In this setting, the social learning dynamics, and hence the demand, is affected by the interplay of two unique features: substitution and ranking effects. The platform can influence the social learning dynamics by adjusting the ranking of the products to ultimately maximize the revenue collected from commission fees for sold items. To formulate the problem in a tractable form, we use a large-market (fluid) approximation and show that consumers eventually learn the products’ quality and characterize the speed of learning. Armed with this backing, we formulate the platform’s ranking problem in the fluid setting, where we assume the perspective of an uninformed platform that does not know the true quality vector but rather learns it through consumers’ review process. We compare different ranking policies based on the worst-case regret with respect to a fully informed platform benchmark. Our analysis yields three main insights. First, a greedy policy that maximizes immediate revenue by displaying products based on current ratings may incur highly suboptimal worst-case regret, as it may relegate the most profitable products to the lowest positions in the ranking if their current rating is not high enough. Second, a simple variant of the greedy policy can sufficiently alleviate the regret by balancing the trade-off between exploration and exploitation. Third, we characterize the critical level of search cost for which the regret does not grow exponentially with the number of products.
- 1 July 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (4) , 1136-1153
Product Ranking in the Presence of Social Learning
This paper studies product ranking mechanisms of a monopolistic online platform in the presence of social learning. The products’ quality is initially unknown, but consumers can sequentially learn it as online reviews accumulate. A salient aspect of our problem is that consumers, who want to purchase a product from a list of items displayed by the platform, incur a search cost while scrolling down the list. In this setting, the social learning dynamics, and hence the demand, is affected by the interplay of two unique features: substitution and ranking effects. The platform can influence the social learning dynamics by adjusting the ranking of the products to ultimately maximize the revenue collected from commission fees for sold items. To formulate the problem in a tractable form, we use a large-market (fluid) approximation and show that consumers eventually learn the products’ quality and characterize the speed of learning. Armed with this backing, we formulate the platform’s ranking problem in the fluid setting, where we assume the perspective of an uninformed platform that does not know the true quality vector but rather learns it through consumers’ review process. We compare different ranking policies based on the worst-case regret with respect to a fully informed platform benchmark. Our analysis yields three main insights. First, a greedy policy that maximizes immediate revenue by displaying products based on current ratings may incur highly suboptimal worst-case regret, as it may relegate the most profitable products to the lowest positions in the ranking if their current rating is not high enough. Second, a simple variant of the greedy policy can sufficiently alleviate the regret by balancing the trade-off between exploration and exploitation. Third, we characterize the critical level of search cost for which the regret does not grow exponentially with the number of products.