Chapter 4: Information systems and Frameworks

Title: Development of an Ontology for Bio-Inspired Design using Description Logics

Author(s): Sungshik Yim, Jamal O. Wilson, David W. Rosen

Address: The George W. Woodruff School of Mechanical Engineering Georgia Institute of Technology, Atlanta, GA 30332-0405, USA | The George W. Woodruff School of Mechanical Engineering Georgia Institute of Technology, Atlanta, GA 30332-0405, USA | The George W. Woodruff School of Mechanical Engineering Georgia Institute of Technology, Atlanta, GA 30332-0405, USA

Reference: International Conference on Product Lifecycle Management 2008 pp. 319 - 328

Abstract/Summary: In this paper, we present an ontology used to capture, retrieve, and reuse novel bio-inspired design solutions and their associated physical architectures, behaviors, functions, and strategies. By adapting these biological design solutions, nature's technology can be leveraged in the design of novel engineering systems. Primarily due to the lack of cross-domain knowledge and a differing functional language, identification of relevant biological solutions by design engineers is a difficult challenge. This challenge can be overcome through using a case-based approach, whereby biological solutions are stored in a biological repository and are related to example engineering solutions. Using the proposed ontology, the designer can store and efficiently retrieve previously stored biological and/or engineering solutions, based on desired functions, architectures, or behaviors. Therefore, the biological repository enables design knowledge capture and reuse in a distributed environment for biomimetics. The proposed ontology is encoded using a Description Logic (DL) known as ALCHI. Description logics are a subset of first-order logic that have been used for information modeling in several application areas, including engineering information management. They are used typically to construct classification hierarchies that can be efficiently searched. We demonstrate the capability of our DL model by: 1) demonstrating that the classification hierarchies that are computed match our biomimetic ontology and 2) demonstrating the retrieval capability using a prototype implementation of our biomimetic ontology.

Order a copy of this article