Application of analogical planning to engineering design

Author

Cook, Diane J.

Author_Institution

Beckman Inst. for Adv. Sci. & Technol., Illinois Univ., Urbana, IL, USA

fYear

1990

fDate

5-9 May 1990

Firstpage

244

Abstract

The analogical graph matching (ANAGRAM) system, which performs analogical planning using a graph-match technique, is described. Given a target problem specification, ANAGRAM finds a similar problem from the database from which a solution can be derived. The difficulties that arise when an analogy fails because the base case is only partially applicable to the current problem are addressed. ANAGRAM offers a solution to this problem by merging the graphs representing several similar base cases, resulting in a virtual base graph that generalizes the individual cases enough to cover a target problem. Three examples that illustrate the use of analogy planning and graph merging in the domain of automatic programming are presented. The techniques described can similarly be applied to tasks in a variety of complex domains

Keywords

CAD; automatic programming; engineering computing; graph theory; pattern recognition; ANAGRAM; analogical graph matching; analogical planning; analogy planning; automatic programming; database; engineering design; graph merging; graph-match technique; target problem specification; virtual base graph; Automatic programming; Databases; Design engineering; Knowledge engineering; Machine learning; Machine learning algorithms; Merging; Technology planning; Urban planning;

fLanguage

English

Publisher

ieee

Conference_Titel

Artificial Intelligence Applications, 1990., Sixth Conference on

Conference_Location

Santa Barbara, CA

Print_ISBN

0-8186-2032-3

Type

conf

DOI

10.1109/CAIA.1990.89196

Filename

89196