Group Technology (GT) is a methodology that utilizes the similarity am
ong parts or machine features. As such, the relevant features of the p
arts have to be considered in order to get the best grouping. This inf
ormation is required in advance and has to remain constant. Contempora
ry manufacturing, on the other hand, advocates flexibility with less c
ommitment to the features used for clustering. This paper analyzes the
affect that information missing or arriving at a different order has
on the grouping of parts. In order to evaluate the variations in infor
mation availability, the paper presents a grouping methodology that al
lows parts to be grouped as they arrive, in a piecewise manner. This a
rticle also compares the quality of the grouping when all the informat
ion is available at the beginning, to the quality of the grouping when
the information is received gradually. The grouping methodology prese
nted in this paper is distributed and asynchronous, utilizing intellig
ent agents. It is implemented using C++ on a PC computer. (C) 1998 Els
evier Science Ltd. All rights reserved.