作者: Ian Frank , David Basin
DOI: 10.1016/S0004-3702(97)00082-9
关键词: Artificial intelligence 、 Normal-form game 、 Game theory 、 Extensive-form game 、 Machine learning 、 Bayesian game 、 Repeated game 、 Combinatorial game theory 、 Computer bridge 、 Complete information 、 Computer science
摘要: Abstract We examine search algorithms in games with incomplete information, formalising a best defence model of such based on the assumptions typically made when information problems are analysed expert texts. show that equilibrium point strategies for optimal play exist this model, and define an algorithm capable computing strategies. Using as reference we then analyse architectures have been proposed game Bridge. These select by analysing some statistically significant collection complete sub-games. Our allows us to clearly state limitations producing analysis, precisely formalise distinguish lead sub-optimality. illustrate these simple trees actual situations from Bridge itself.