The life span of facilities produced by the Architecture-Engineering-Construction industry is typically 25 years or more Several distinct phases characterize the life span of a facility. Each of these phases involve numerous participants from different professional disciplines. These participants generate and use a lot of information about the facility. Current methods used by the industry to convey this information are drawings and specifications. However. these drawings and specifications reflect only a summary of the information generated and used by the project participants This summarized information only describes the product. Information about the process of generating these information becomes implicit in the drawings and specifications. Rationale is the collective term for this set of implicit process information.

The main issue addressed by this dissertation is the need to communicate design rationale information. Design rationale is a subset of the entire rationale generated for a facility Design rationale refers to information about the design process. Explicitly stating design rationale information reduces the chance of misinterpreting design drawings and specifications.

The primary objective of this dissertation is to determine a data structure capable of representing design rationale information. This data structure also allows a computer system to perform analytical tasks on the design rationale data. Examples of analytical tasks a computer system can perform on design rationale data include: generating a parameter dependency network and resolving data conflicts. This dissertation defines this data structure as two separate but complementary modules. The Knowledge Representation Module assists in gathering project-specific product information. The Rationale Storage Module assists in capturing project-specific process information. This dissertation discusses each of these two modules in detail.

The secondary objectives of this dissertation include: (1) defining a computer program architecture, (2) creating a computer program interface, and (3) verifying the appropriateness of the data structure in representing design rationale. A proof-of-concept computer program, DRIVE, applied to an actual value engineering study project accomplishes these objectives.