Bayesian Belief Networks for Diagnosing Asphalt Paving Construction Problems

摘要

The construction of a knowledge-based Bayesian belief network to provide decision support for asphalt paving construction. The HUGIN software was employed to develop the system. The model consists of three layers of nodes. The lop layer consists of nodes describing observable asphalt paving problems and the bottom layer consists of problem diagnoses and corrective actions. Through a user interface, model users can input observed problems by changing probabilities in the network. The changes are propagated through the network, and users can observe changes in the diagnoses probabilities.