This paper describes an integrated approach for design evaluation and partner selection for agile manufacturing. The information requirements for agile manufacturing are integrated product models that represent a product design and support design evaluation functions, manufacturing process models that enable high-level process planning, and manufacturing plant models that describe the specific capabilities and performance of a candidate partner. This approach utilizes two forms of partner-specific design evaluation: the retrieval of similar designs and high-level process planning. Partner selection results from these evaluation methods.
This system provides an agile manufacturing firm with useful information for forming the partnership (virtual enterprise) and therefore has two significant benefits: the firm selects partners based on their ability to manufacture the specific product being designed (not the manufacturing plant's general performance), and the firm receives feedback about the design at an early point in the design stage---before the firm selects the partners. The system was designed for use by a designer within a manufacturing firm that participates in virtual enterprises. In addition, the system can be used by a broker whom such a manufacturing firm hires to select partners for a virtual enterprise.
We have implemented the system using object-oriented programming and database technologies. The information models were developed in the EXPRESS modeling language and used to define the schemas of ObjectStore object-oriented databases [28]. The system modules comprise C++ code and routines from the ObjectStore database management language. The user interface of the system uses a library of SUIT functions for entering and displaying data [7]. The CAD system uses the ACIS [33] solid modeling kernel. ACIS is also used for geometric reasoning. The system uses HOOPS [16] for product display.
The primary contribution of this work is the use of design evaluation as a method for partner selection, facilitated by partner-specific design evaluation techniques. This research has developed novel information models (the integrated STEP-based product model, the OOGT information model, and the manufacturing resource model) and decision support methods for the search and sort of similar products, the generation and evaluation of high-level process plans, and the selection of the optimal manufacturing alternative. The system development includes the object-oriented implementation of a STEP-based product model, the construction of methods for translating product data between the geometric and topological models of STEP and ACIS, and methods for automatically generating GT codes.
This work was sponsored by the U.S. Army Tank-Automotive Command under grant DAAE-07-93-C-R086. The research team was led by University of Maryland researchers and included researchers from the State University of New York at Buffalo, the National Institute for Standards and Technology, Westinghouse Electric Corporation, and Martin-Marietta. The researchers at the University of Maryland were partially supported by the Institute for Systems Research under grant NSFD CD 8803012.