Abstract
Multiply-sectioned Bayesian networks (MSBNs) provide a coherence framework for multi-agent distributed interpretation tasks. Inference in MSBNs can be performed in a time-sharing environment or a distributed environment. Inference in a fully distributed environment not only improves the efficiency with respect to the time-sharing counterpart but is also necessary for problem domains that are physically distributed. Although evidential inference has been implemented in a time-sharing fashion, new issues must be resolved in order to perform such inference in a fully distributed environment. In this paper, we present these new issues and our solutions. We also demonstrate our implementation experimentally.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have