BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:The alpha-Potential Game Paradigm: Theory\, Algorithms\, and Appli
 cations - Xinyu Li (University of Oxford)
DTSTART:20251110T144000Z
DTEND:20251110T152000Z
UID:TALK238432@talks.cam.ac.uk
DESCRIPTION:Designing and analyzing non-cooperative multi-agent systems th
 at interact within shared dynamic environments is a central challenge acro
 ss many established and emerging applications\, including autonomous drivi
 ng\, smart grid management\, and e-commerce. A key objective in these syst
 ems is to identify Nash equilibria\, where no agent can benefit by unilate
 rally deviating from its strategy. However\, computing such equilibria is 
 generally intractable unless specific structural properties of theinteract
 ions can be leveraged.Recently\, we have developed a new paradigm known as
  the alpha-potential game framework for studying dynamic games. This talk 
 illustrates the framework through a class of dynamic games motivated by ga
 me-theoretic models of crowd motion. We show that analyzing alpha-Nash equ
 ilibria reduces to solving a finite-dimensional control problem. Beyond pr
 oviding viscosity and verification characterizations for general games\, w
 e examine in detail how spatial population distributions and interaction r
 ules shape the structure of alpha-Nash equilibria\, in particular for crow
 d motion games.\nTheoretical insights are complemented by numerical experi
 ments based on policy gradient algorithms\, which highlight the computatio
 nal advantages of the alpha-potential game framework for efficiently compu
 ting Nash equilibria in dynamic multi-agent environments.&nbsp\;
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR