BEGIN:VCALENDAR VERSION:2.0 PRODID:ILLC Website X-WR-TIMEZONE:Europe/Amsterdam BEGIN:VTIMEZONE TZID:Europe/Amsterdam X-LIC-LOCATION:Europe/Amsterdam BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 TZNAME:CEST DTSTART:19700329T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 TZNAME:CET DTSTART:19701025T030000 RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT UID:/NewsandEvents/Archives/2010/newsitem/3616/4-N ovember-2010-Empirical-Game-Theoretic-Analysis-for -Practical-Strategic-Reasoning-Michael-Wellman DTSTAMP:20101028T000000 SUMMARY:Empirical Game-Theoretic Analysis for Prac tical Strategic Reasoning, Michael Wellman ATTENDEE;ROLE=Speaker:Michael Wellman DTSTART;TZID=Europe/Amsterdam:20101104T110000 DTEND;TZID=Europe/Amsterdam:20101104T000000 LOCATION:Room L016 ('Hypathia'), CWI, Science Park 123, Amsterdam DESCRIPTION:The games agents play - in markets, co nflicts, or most other contexts - often defy stric t game-theoretic analysis. Games may be unmanageab ly large (combinatorial or infinite state or actio n spaces), and present severely imperfect informat ion, which could be further complicated by partial dynamic revelation. Moreover, the game may be spe cified procedurally, for instance by a simulator, rather than in an explicit game form. With colle agues and students over the past few years, I have been developing a body of techniques for strategi c analysis, adopting the game-theoretic framework but employing it in domains where direct "model-an d-solve" cannot apply. This empirical game-theoret ic methodology embraces simulation, approximation, statistics and learning, and search. Through appl ications to canonical auction games, and rich trad ing scenarios, we demonstrate the value of empiric al methods for extending the scope of game-theoret ic analysis. This perspective also sheds insight i nto behavioral models and bases for predicting joi nt action in complex multiagent scenarios. For mo re information, contact H.van.Hasselt at cwi.nl X-ALT-DESC;FMTTYPE=text/html:\n
The game s agents play - in markets, conflicts, or most oth er\n contexts - often defy strict game-theo retic analysis. Games may\n be unmanageably large (combinatorial or infinite state or\n action spaces), and present severely imperfect i nformation,\n which could be further compli cated by partial dynamic\n revelation. More over, the game may be specified procedurally,\n for instance by a simulator, rather than in a n explicit game\n form.
\n\n With colleagues and students over the past few years, I have\n been developing a body of techniques for strategic analysis,\n ado pting the game-theoretic framework but employing i t in\n domains where direct "model-and-solv e" cannot apply. This\n empirical game-theo retic methodology embraces simulation,\n ap proximation, statistics and learning, and search. Through\n applications to canonical auctio n games, and rich trading\n scenarios, we d emonstrate the value of empirical methods for\n extending the scope of game-theoretic analysi s. This\n perspective also sheds insight i nto behavioral models and\n bases for predi cting joint action in complex multiagent\n scenarios.
\n \nFor more informat ion, contact H.van.Hasselt at cwi.nl
URL:/NewsandEvents/Archives/2010/newsitem/3616/4-N ovember-2010-Empirical-Game-Theoretic-Analysis-for -Practical-Strategic-Reasoning-Michael-Wellman END:VEVENT END:VCALENDAR