Data processing: artificial intelligence – Fuzzy logic hardware – Defuzzification processing
Reexamination Certificate
2000-03-03
2004-07-13
Starks, Jr., Wilbert L. (Department: 2121)
Data processing: artificial intelligence
Fuzzy logic hardware
Defuzzification processing
C706S008000
Reexamination Certificate
active
06763337
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The following invention relates to determining and analyzing fuzzy conceptual design alternatives, and specifically to applying a volumetric defuzzification approach in conceptual design.
2. Related Art
Modern companies pursue ever increasingly innovative and thoughtful approaches to generating design alternatives that fit customer needs. This applies across many types of companies, producing many types of products, from intellectual products like software, to manufacturing of cars. Decision making must be responsive to customer needs, and strategic, from conceptual system design, to preliminary system design, and finally to a detailed system design leading to production and construction.
For conceptual design, the evaluation methodology includes a number of strategic decisions and applies to the many processes of any given company, including but not limited to manufacturing distribution, deployment, installation, operations, sustaining maintenance, support and disposal.
A noted modern approach to conceptual design uses Pugh's approaches. (See S. Pugh, “The application of CAD in relation to dynamic/static production concepts,” Proc Int Conf Eng Des, Copenhagen, August 1983, pp. 564-571; S. Pugh, “Total design: Integrated methods for successful product engineering,” Addison-Wesley, New York, 1991.) Evaluation criteria are delineated based on need analysis, aided by design methods. The initial design is subjected to a “controlled convergence” process, facilitated by a concept selection matrix.
Pugh's method analyzes the customer's need and generates a requirements definition. A crisp, exact feasible design envelope is created, within which potential system design concepts are generated. From the potential design concepts, a conceptual design selection matrix is then generated. The potential design concepts consist of a number of identified, crisp variables, called Design Dependent Parameters (DDPs). The DDPs are application specific, since the design pertains to the particular technological or industrial process, and the items being investigated pertain to one or more items within that process. Exemplary DDPs are reliability, maintainability, environment compliance, visual quality, and safety.
A number of modem techniques have applied fuzzy techniques to Pugh's crisp solutions. Instead of using crisp definitions for DDPs, fuzzy, linguistic, more human definitions are applied. For example, a crisp definition for the reliability of a machine can be whether the mean time before failure (MTBF) is “40,000 hours.” On the other hand, a fuzzy definition for the reliability MTBF can be “approximately 40,000, ” where the meaning of the set “approximately 40,000” depends upon how fuzzy (or alternatively crisp) the linguistic definition is meant by the user. For example, for a very fuzzy set, even the value 80,000 may have a non-zero profile value associated with it, such that it is covered under the umbrella of “approximately 40,000.” On the other hand, for a less fuzzy (more crisp) set, “approximately 40,000” may have non-zero profile values only between 39,900 and 40,100.
After the DDPs are identified, defuzzification techniques must be applied to compare fuzzified variables. Suppose the DDP in question is the reliability parameter. The fuzzy value for the anticipated reliability profile (the reliability anticipated by the product designer) is compared to the reliability requirement profile (the reliability required by the customer's needs). Existing defuzzification approaches are not applicable to this problem since the objective in this case is not to create an ordering between the fuzzy profile, but rather to compare them from the perspective of compliance. Accordingly, none of the previous techniques have properly accounted for sensitivity, meaning that when the input is mapped to the output, the difference between the results from the various inputs is non-trivial in nature. What is required is an exact, more satisfactory way to defuzzify the parameters, in addition to providing sensitivity analysis.
SUMMARY OF THE INVENTION
The present invention is directed to a method, and a system for employing the method, for determining compliance for a feasible design dependent parameter (DDP) between an anticipated DDP fuzzy profile and a required DDP fuzzy profile. The method includes (a) creating a weighted wedge by projecting an image of a surface common to the anticipated DDP fuzzy profile and the required DDP fuzz profile at an angle from the surface; and (b) finding an overlap volume of the weighted wedge between a projection of the anticipated DDP fuzzy profile onto the image and a projection of the required DDP fuzzy profile onto the image.
The angle can be calculated in a manner to make the overlap volume greater if any one of the anticipated DDP fuzzy profile and the required DDP fuzzy profile have projections with higher respective preference levels. The method can also include generating a feasibility index for the DDP by dividing the overlap volume by a volume obtained from the projection of the anticipated DDP fuzzy profile.
The common surface can be a linear projection plane. The common surface can also be a non-linear projection plane designed to provide relatively greater differences between one or more of the overlap volumes at higher respective preference levels for any one of the anticipated DDP fuzzy profiles and the required DDP fuzzy profiles.
The method can be part of larger system, which includes the following processes: (a) determining a feasible design space for a design concept by finding one or more DDPs for the concept; (b) determining one or more feasible DDPs for the design space, including applying the above creating step and finding step one or more times; and (c) rating and ranking feasible design concepts by consolidating relative priorities of the one or more feasible DDPs.
The invention will be understood by those skilled in the relevant art from the descriptions provided below, though various changes in form and details may be made without departing from the spirit and scope of the invention.
REFERENCES:
patent: 5446827 (1995-08-01), Shigeoka et al.
patent: 5546501 (1996-08-01), Yoshitake et al.
patent: 5561739 (1996-10-01), Muraji
patent: 5751908 (1998-05-01), Madau et al.
patent: 5778149 (1998-07-01), Eichfeld
patent: 5852708 (1998-12-01), Shyu et al.
patent: 5940814 (1999-08-01), Jiang et al.
Verma, D., “A Fuzzy Set Paradigm For Conceptual System Design Evaluation”, Virginia Polytechnic Institute and State University, Doctorial Dissertation, Dec. 1994.*
Verma et al.; “Systematically Identifying System Engineering Practices and Methods”. IEEE Transactions on Aerospace and Electronic Systems, vol. 33, No. 2, Apr. 1997, pp. 587-595.*
Verma et al.; “Application of Fuzzy Logic In The Assurance Sciences”. 1994 Proceedings Annual Reliability and Maintainability Symposium, Jan. 1994, pp. 436-441.*
Verma et al.; “Analyzing a Quality Function Deployment (QFD) Matrix: An Expert System Based Approach to Identify Inconsistencies and Opportunities”. Journal of Engineering Design, vol. 9, Iss 3, Dec. 1998, pp. 1-9.
Smith Caroline
Verma Dinesh
Aitken Andrew C.
Booker Kelvin E
Lockheed Martin Corporation
Starks, Jr. Wilbert L.
Venable LLP
LandOfFree
Weighted wedge defuzzification for conceptual system design... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Weighted wedge defuzzification for conceptual system design..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Weighted wedge defuzzification for conceptual system design... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3218287