Use of Knowledge Graphs to Formalize Decisions in Preserving Pavements

Abstract

A methodology is presented that utilizes the concept of knowledge graphs to formalize the decision processes involved in pavement preservation. The methodology addresses the problem‐solving processes of these ill‐structured decision tasks: symptom observation, condition diagnosis, and treatment identification. A decision framework is presented in the form of knowledge graphs and involves problem decomposition in taxonomic hierarchies, term interpretation of the associated contexts, and heuristics organization. The defined knowledge graphs, coupled with a frame representation scheme, constitute the basis for developing a prototype knowledge‐based system that embodies expertise on pavement preservation at the New York State Thruway Authority. Important aspects of formalizing the problem‐solving processes and acquiring domain knowledge are evaluated and discussed. It is concluded that the developed methodology, which clearly identifies the structure of knowledge and the relation of contexts, contributes significantly to the design of knowledge‐based systems.