Projects
Imperfect Information Games; Models and Analyses (InIGMA)
Phd. Project
September 2002 - 2006Prof. Dr. J.F.A.K. van Benthem (promotor)
Dr. P. van Emde Boas (supervisor)
Sjoerd Druiven
Merlijn SevensterIn contemporary logic efficiency has become as much an issue as effectivity used to be in the past. Problems, such as model checking - given a model and a formula from a logical language, how much effort does it take to conclude that the formula holds or does not hold in the model? - are already taught in the introductory courses on logic. Conversely, problems in logic have been recognized to be representative of standard degrees of tractability and intractability in complexity theory. On the other hand, the topics of logic and game theory get more and more intertwined, viz. evaluation games - played according to a formula; it's winning conditions determined by the truth of that formula, w.r.t. some model. Consequently, it's no miracle that the intersection of complexity theory, logic and game theory is tudied. In this field questions arise like: Given an evaluation game, what's the complexity of computing the power of some player to always win the game? For games in which information is perfectly known by all players - like chess -, an extensive amount of literature has been produced during the last 30 years. On the other hand, in the case of imperfect information games - like most card-games -, complexity questions have been left mainly unposed. In the InIGMA-project we indeed will pose and investigate complexity questions considering games of imperfect information.
Knowledge Development in Games of Imperfect Information
Msc. Project ps pdf
June 2001 - May 2002Sjoerd Druiven
H.H.L.M. Donkers (supervisor)
Dr.ir. J.W.H.M. Uiterwijk (supervisor)
Dr. L.C. Verbrugge (supervisor)In a game of perfect information, players are fully informed about the current game position. In games of imperfect information, players have incomplete information about the current game position. The development of knowledge in games of imperfect information is studied. The aim is not only to model the development of knowledge about game states, as is done by current dynamic epistemic logics, but also to model knowledge which is a result of strategic information. These forms of knowledge have up to now not been modeled in literature on knowledge in games.
By describing games in dynamic-epistemic logic, knowledge and knowledge development of players who have strategic information about opponents can be modeled.
Valuatie van argumenten in frames en het aanmaken van geprefereerde en stabiele extensies
June - August 2000
Draft ps pdf
Sjoerd Druiven
Maarten Grachten
Dr. L.C. Verbrugge (supervisor)Argumentatie frameworks zijn modellen die formele discussies modelleren. Een framework bestaat uit een verzameling argumenten en een binaire relatie gedefineerd op deze argumenten, een zogenaamde aanvalsrelaties. Er bestaan dus argumenten, waarvan de interne structuur niet is gedefineerd, en argumenten die andere argumenten aanvallen. Bepaalde verzamelingen van argumenten hebben een specifieke betekenis, welke kunnen worden gezien als standpunten. In samenwerking met mede-student Maarten Grachten is een algoritme geschreven wat deze standpunten gegeven een argumentatie framework kan uitrekenen.
-
June - Juli 1999
Sjoerd Druiven
Maarten Grachten
Dr. L.C. Verbrugge (supervisor)Het spel MastersMinds is een variant op Mastermind, waarbij twee spelers beide een geheime te kraken code kiezen. De code wordt gekraakt door het stellen van vragen, gelijk zoals mastermind. Het interressante aspect van het spel is dat de vragensteller ook zijn eigen vragen moet beantwoorden. Mede-student Maarten Grachten en ik hebben een algoritme geschreven dat het spel kan spelen en winnen.
(none) hits in May 2012