抄録
Marine Engineers acquire knowledge empirically from monitoring data and checking the conditions of marine engine plants. They make the best use of their knowledge in order to manage their job safely, efficiently and economically. On the other hand, the development of communication technology now allows shore-side managers to collect monitoring data from several marine engine plants. Using this data, shore-side managers gather information which can be useful for supporting ships. However, these useful methods have not yet been completely developed. In this research, our aim is to construct an artificial intelligence system called “Virtual Ship Engineer” or VSE, that can accumulate knowledge based on its own autonomous learning processes in way similar to how marine engineers do. The authors divided this learning process into three phases namely, “Data Predicting Model”, “Knowledge Evaluating Model” and “Knowledge Updating Model” and constructed VSE’s autonomous learning model. In this paper, the authors explain the details of these three models using fuzzy reasoning and the steepest descent method and two simulations by using monitoring data to verify the validity and possibility of VSE’s autonomous learning model.
