MLSE Seminar

Author: Christopher Kops (UM,KE)
Date and time: December 12th, 2023 (13:00-14:00), TS53 A0.24
Title: Who Influences Whom? Identifying the Social Network from Behavior
Abstract: ​Individual behavior is often influenced by that of peers. The identification of such peer influence from observable behavior constitutes a classic problem in the social sciences. When can we identify the social network of influence? We propose a simple model of social influence and interactions. In this framework, we characterize exactly which behavior and graph structures are informative when it comes to the question of identifying the underlying network of influence.

Author: Wout de Vos (Tilburg University)
Date and time: November 28th, 2023 (13:00-14:00), TS53 A1.23
Title: Influencing opinion networks – optimization and games
Abstract: We consider a model of influence over a network with finite-horizon opinion dynamics. The network consists of agents that update their opinions via a trust structure as in the DeGroot dynamics. The model considers two potential external influencers that have fixed and opposite opinions. They aim to maximally impact the aggregate state of opinions at the end of the finite horizon by targeting with precision one agent in one specific time period. In the case of only one influencer, we characterize optimal targets on the basis of two features: shift and amplification. Also, conditions are provided under which a specific target is optimal: the maximum-amplification target. In the case of two influencers, we focus on the existence and characterization of pure strategy equilibria in the corresponding two-person strategic zero-sum game. Roughly speaking, if the initial opinions are not too much in favour of either influencer, the influencers’ equilibrium behaviour is also driven by the amplification of targets.

Author: Günnur Ege Bilgin (Bonn University)
Date and time: October 31st, 2023 (13:00-14:00), TS53 A1.23
Title: Decentralized Many-to-One Matching with Random Search
Abstract: : I study a decentralized many-to-one search model with frictions, where a finite number of firms and workers meet randomly and time is nearly costless. Different than the existing literature, I show that stable matchings of the many-to-one market may not be enforceable under Markovian search equilibria even if agents are sufficiently patient. Nevertheless, they are enforceable in markets with aligned preferences, and I construct the equilibrium strategy profile that enforces the unique stable matching. Moreover, in many-to-one search, firms collect workers in a cumulative manner, where they expand their employment set over time. This distinct characteristic affects the search process fundamentally by allowing firms to strategically use the number of their vacancies over time, leading to more favorable outcomes for themselves. Therefore, unlike centralized markets, there is no straightforward related-market connection between one-to-one and many-to-one search models in the decentralized case. I establish one sufficient condition on the preferences of agents to retrieve the connection, which also prevents the enforcement of unstable matchings in equilibrium.

Author: Joep van Sloun (UM,KE)
Date and time: October 10th, 2023 (13:00-14:00), TS53 A0.24
Title: Rationalizable Behavior in the Hotelling Model with Waiting Costs
Abstract: : This paper revisits the Hotelling model with waiting costs. For the two firm case and asymmetric efficiency levels and the three firm case with symmetric efficiency levels, there is no Nash equilibrium in pure choices. That is why we use a different solution concept, called point rationalizability. We characterize the set of point rationalizable choices for the two firm case, and hopefully more results will follow for three and more firms in the future.

Author: Shaul Rosner (Reichman University).
Date and time: September 29th, 2023 (13:00-14:00), TS53 C-1.05
Title: Congestion Games with Rank-Based Utilities
Abstract: : Studies in behavioural science show that individuals are often concerned primarily about their relative welfare, rather than their absolute well-being. In this talk we will define and study a variant of congestion games that reflects this phenomenon. In a congestion game with rank-based utilities, the players are partitioned into competition sets, and the goal of every player is to minimize its cost relative to its competitors. Specifically, the primary goal of a player is to minimize the rank of its cost among its competitors, while minimizing the cost itself is a secondary objective. This fits environments with strong competition among the participants, in which the relative performance of the players determine their welfare. We show that games with rank-based utilities are significantly different from classical congestion games, and that competition may lead to a poor outcome with respect to the existence of Nash Equilibrium and its quality. We consider two classes of congestion games: Job-scheduling games, in which each player selects a machine, aiming to have a low completion time relative to its competitors, and cost-sharing games, in which players fairly share the activation cost of resources they use, each aiming to have a relatively low payment.

As we prepare for the upcoming spring semester in 2024, please reserve the following dates in your calendars:

30th of January     
6th of February     
19th of February