M.S. Thesis Presentation by Brian P. Corbett
Wednesday, January 29, 2003

(Dr. David Rosen, advisor)

"Configuration Design Methods and Mathematics for Product Families"

Abstract

Product families, groups of products that share a common underlying structure,
have recently been adopted in industry to help companies meet the demands of
diverse customers in a competitive global marketplace, lessen the time needed
to develop new products, and reduce costs by sharing common components among
many products. Design of product families often involves the use of a product
platform: a group of components, modules, interfaces, and technologies that are
shared among all members of the product family. Configuration design serves as
a useful method for identifying potential platforms for a product family.
Configuration design involves the identification of components in a product and
a description of the relationships and physical arrangement of those components.

To solve configuration design problems in a systematic manner, discrete
mathematical models based on set and graph theory are used. Four viewpoints
describing modularity, physical connectivity, assembly, and relations involving
energy or material flows in the product are modeled using discrete mathematics
and are utilized to solve configuration problems. Using these mathematical
models as a framework, primary research contributions were made in the
following areas:

1) Mathematical modeling of flow relations in a product from a configuration
perspective
2) Development of computational models for combining multiple configuration
viewpoints
3) Identification of a method for partitioning the configuration design process
so that problems involving products containing a large number of components may
be solved

The computer implementation of configuration design methods was also undertaken
in order to make the solution of configuration problems practical. Algorithms
for principal configuration design tasks are outlined in the thesis. The
configuration methods developed are tested with three case studies: a
flashlight, a mechanical pencil, and an automotive body structure. In the
flashlight case, application of the flow viewpoint allows product
configurations to be more accurately modeled. The mechanical pencil case is
used to investigate the efficiency of computational models for combining
multiple configuration viewpoints. The automotive body structure case
demonstrates how partitioning the configuration design problem enables larger
problems to be solved within the limits of available computing

power.