Evidence-driven decision support in critical infrastructure management through enhanced domain knowledge modeling

Abstract

Effective critical infrastructure management in dynamically changing service environments requires understanding and inferring unknown knowledge from complex heterogeneous dataset to reason about multi-dimensional complex problem solving activities by aggregating supporting evidences. While the attributes of the database table only describe data and certain notions from the database relational schema, they do not describe the higher-level concepts or the knowledge from the domain that are commonly thought of and referred by engineers who need to inspect and manage the infrastructure with a holistic viewpoint. Thus, engineers have to work with rudimentary data-level attributes that, further, complicates the critical infrastructure management, which essentially needs efficient, effective, and informed decision making. Ontology enables to solve a complex problem where the underlying domain concept provides collective understanding of the data based on the domain knowledge from multi-dimensional resources. Enhanced domain knowledge modeling is applied for transportation infrastructure asset management that requires bridge inspectors to make decisions based on complex multi-layered heterogeneous data, such as, infrared image data, aerial photo data, ground-mounted LIDAR data, etc. The ontological concepts represent the process knowledge and assessment knowledge and it will be further used to support the bridge inspectors and their inspection process, whereas data are the ground facts. This process knowledge plays an important role to bridge the ground facts and the high-level concept space and provides the mapping of the complex data space to the easily comprehensible conceptual space. In making critical decisions, these become crucial evidences in justifying decisions made as well as in making uniform decisions among different subject matter experts through the common understanding.