Quantitative Discovery and Reasoning about Failure Mechanisms in Pavement

This paper describes a hybrid system for quantitative discovery and its related qualitative reasoning. The system HOTEP, is currently under development at MIT. The application area is the Failure Mechanisms in Pavement. The separate areas of Quantitative Discovery, Qualitative Reasoning, and Explanation-based generalization have been investigated by other researchers, but no attempt has been made to hybridize all of them together into a robust discovery system. Faced by a real need for knowledge in the evolving domain of the Failure mechanisms in pavement, we found an incentive to pursue the research in this direction. The paper emphasizes the importance of integrating mathematical induction tools with symbolic techniques to develop better understanding of the new field. Beside the discovery activities, the system is capable of handling competing theories. The quantitative discovery system in HOTEP has been implemented already, and it was tested in several areas. We propose harvesting knowledge bases from the sites so we can combine experience, then broadcast said synthesis back to the sites.