Knowledge Processing for a Construction Management Database

The objectives of project control systems go beyond documentation to recognition of problems and evaluation of their causes. Through the use of broad information systems, a construction manager today is able to assemble a wide variety of data about his projects or programs. Retrieval, analysis and interpretation of the meaning of these data, however, usually requires the user to have detailed knowledge about the structure and content of the database, use of a data retrieval language, and have further programming skills to have the computer perform analyses. The work reported in this thesis has three goals: to ease retrieval with near-natural language query capabilities, to acquire and accumulate user defined knowedge, and to exercise such knowledge for translation from English terms to database fields or their values, access to database or other information sources, or pre-access and post-access processing. A knowledge based query system, called Expert-MCA, has been developed by coupling the techniques of natural language processing, knowledge-based expert systems, and database management systems. The system architecture will be presented, followed by a description of how knowledge is represented and processed in answering English-like queries. An object-oriented formalism is applied to represent knowledge required in the process of reasoning. Expert-MCA's reasoning capability has been extended by using context driven reasoning, a generalization of reasoning mechanisms used in framed-based systems. Thesis Supervisor: Dr. Robert D. Logcher Title: Professor of Civil Engineering