Diagnosis from Bayesian Networks with Fuzzy Parameters-A Case in Supply Chains

Abstract

Bayesian networks have been widely used as knowledge models in business, engineering, biomedicine, and so on. When a network is learned with incomplete knowledge, the numerical model based on probability theory needs to be extended. This study presents a robust approach for diagnosis from Bayesian networks with fuzzy parameters. A simulation algorithm is designed to answer queries from the graphical models. The formulation of piecewise linear possibility distribution functions maintain the scalability in exact approaches.