AN AI-BASED DECISION SUPPORT SYSTEM FOR NAVAL SHIP DESIGN

Naval ship design is a complex, iterative decision-making process that requires not only qualitative reasoning and quantitative computations, but also effective handling of massive amounts of information. The development of a systematic procedure in the form of an interactive Decision Support System (DSS) incorporating artificial intelligence (AI) techniques can assist in improving the design process. The DSS reported in this paper integrates conventional naval ship design application software with an expert system. Modeling techniques such as a spreadsheet, a database management system, and graphic display capability are incorporated as well. This DSS facilitates "what-if" analysis of design problems and assessment of the trade-offs between performance and cost. The DSS will aid rapid generation of design alternatives and facilitate their evaluation. A final recommendation is made based on multiattribute decision analysis using the straightforward points allocation method and the analytic hierarchy process. During a typical consultation process, the user is prompted for inputs to the naval ship design process and is provided with a wide range of reports in graphical and numerical form. During a structured system evaluation, the DSS displayed considerable accuracy, flexibility, speed, user-friendliness, and cost-effectiveness. The reliability and validity of the system were also confirmed. These evaluation procedures provide a replicable methodology for verification and validation of this type of AI-based decision support system.