Reference: Nayak, P. P.; Joskowicz, L.; & Addanki, S. Automated Model Selection using Context-Dependent Behaviors. Knowledge Systems Laboratory, May, 1991.
Abstract: Effective problem-solving about complex engineered devices requires device models that are both adequate for the problem and computationally efficient. Producing such models requires identifying the relevant device features and determining applicable simplifications. This paper presents a method for automatically constructing a device model by selecting an appropriate model for each of the device’s components using the context in which it operates. We introduce context-dependent behaviors (CDBs), a frame-like component behavior model representation for encapsulating contextual modeling constraints. We show how CDBs are used in the model selection process by exploiting constraints from three sources: the structural and behavioral contexts of the components, and the expected behavior of the device. We describe an implemented program for model selection. The inputs are the structure of the device--the components of the device and structural relations between them--the expected device behavior, and a library of CDBs. The output is a set of component CDBs forming a structurally and behaviorally consistent device model that achieves the expected behavior. We demonstrate the program on a temperature gauge.
Full paper available as ps.