Creating and using models for engineering design: a machine-learning approach

An adaptive and interactive modeling system (AIMS) that integrates simulation, optimization and machine learning to help engineers make design decisions is described. AIMS views engineering decision making as a two-phase process of creating and then using models. The competitive relation learner and the induce-and-select optimizer, AIMS's two main components, and their roles in both phases of decision-making are discussed. AIMS's role in supporting the design of a diesel engine that outputs power within the 440- to 460-kW range and consumes the least amount of fuel is also discussed.<
>