Structured Modeling Language for Automated Modeling in Causal Networks

Abstract

The paper presents a structured modeling lan guage (SML) and a relational database frame-work for specification and automated genera tion of causal models. The framework describes a relational database scheme for encoding a li brary of causal network templates modeling the basic components in a modeling domain. SML provides a formal language for specifying mod els as structured components that can be com posed from the basic components. The lan guage enables specification of models as param eterized relational queries that can be instan tiated for specific model instances. The pa per describes an algorithm that, given a library and a specification, computes a causal model in time and space linear in the number of ba sic components. The algorithm enables model reuse by combining model fragments from the template library to compose new models. The present automated modeling approach has been implemented using the structured query lan guage (SQL) and a relational database envi ronment. The approach has been successfully used for modeling an automated work-cell in a real-life digital manufacturing application. 1