Using Building Data Models to Represent Workflows and a Contextual Dimension

Abstract

The context-workflow relationship is often poorly defined or forgotten entirely. In workflow systems and applications context is either omitted, defined by the workflow or defined based on a single aspect of a contextual dimension. In complex environments this can be problematic as the definition of context is useful in determining the set of possible workflows. Context provides the envelope that surrounds the workflow and determines what is or is not possible. The relationship between workflow and context is also poorly defined. That context can exist independently of workflow is often ignored, and workflow does not exist independently of context. Workflow representations void of context violate this stipulation. In order for a workflow representation to exist in a contextual dimension it must possess the same dimensions as the context. In this thesis we selected one contextual dimension to study, in this case the spatial dimension, and developed a comprehensive definition using building data models. Building data models are an advanced form of representation that build geometric data models into an ob ject-oriented representation consisting of common building elements. The building data model used was the Industry Foundation Classes (IFC) as it is the leading standard in this emerging field. IFC was created for the construction of facilities and not the use of facilities at a later time. In order to incorporate workflows into IFC models, a zoning technique was developed in order to represent the workflow in IFC. The zoning concept was derived from multi-criteria layout for facilities layout and was adapted for IFC and workflow. Based on the above work a zoning extension was created to explore the combination of IFC, workflow and simulation. The extension is a proof of concept and is not intended to represent a robust formalized system. The results indicate that the use of a comprehensive definition of a contextual dimension may prove valuable to future expert systems.