MACC'97 Session: Probabilistic Model in Multiagent Systems
Abduction is a defeasible reasoning and can be applied to various applications including designing and planning. However, the order of the inference speed of it is proved to be NP-complete. Therefore, various parallelism or concurrentism approaches have been studied. Furthermore, abduction is said to be equal to Bayesian model. Probabilistic reasoning model is also useful for models dealing with uncertainty. Recently, Bayesian model is mapped to the market model and the possibility of distributed probabilistic inference has been discussed. This paper will show the framework of multi-agent abduction which makes use of Bayesian network and Game theory, then discuss the merit of it and problems.
multi-agent abduction, game theory, Bayesian model
PS
file(+gzip) (in Japanese)