Decision- Based Design Value Design

Nand K. Jha
Professor of Mechanical Engineering
Manhattan College
Riverdale, New York 10466

The designer is concerned not just with making a good design; he must create the best possible design. By far the the greatest part of the design process is intuitive; the analytical techniques only an aid to the process, not the process itself. These analytical techniques are part of the science of design. They owe much to the theory of mathematical programming.

Throughout the design process the designer is confronted with the problem of uncertainty. These uncertainties may involve risk ,: where each action has several possible outcomes for which the probabilities are known.and pure uncerertainty; where each action has several possible oucomes for which the probabilities are unknown. Desgner never has adequate information for design. In present day highly competitive market, engineering has become more sophisticated and precise, and designer find it increasingly desirable to apply a measure to their uncertainity, so that it becomes more meaningful and consistent. I will also like to talk a little about designer himself or herself. The basic guide lines for all types of design are:

1) Get a clear notion of what you desire to accomplish, then you will probably get it.

2) keep a sharp look-out upon your materials; get rid of every pround of mateial you can do without. Avoid complexities and make everything as simple as possible.

3) remember the get-ability of parts.

Although each step of the design path, design engineers have to apply their own personal design methodology, which they must develop themselves. In general, the method should:

1) Foster creativity 2) Acknowledge the creativeness of others 3) Do not depend on luck or ignore a problem in the hope that it will go away 4) Be disciplined and well organised so the design can be passed onto others for deatailing or completion. 5) Respect simplicity and fundamental knowledge of how and why things work 6) Continually subject designs to value analysis in an effort to reduce cost with an equal or increased level of quality.

A really good design engineer is separated from a part detailer by the former's ability to generate conceptually design from which final designs evolve. An important exercise to help you become better at generating conceptual design is to adapt the philosophy " How does that work?" or "Why does that catch my eye." and apply it to everything you see in daily life

Decision Theory:

Decision theory is concerned with decision making under uncertainty. In design many decisions are made with incomplete knowledge. Evolution has given engineers the intellectual facility of intuition for dealing with uncertainty. However, a process should be developed for rational use of these knowledges. Decision making in design may occur under conditions of (1) Certainty (2) Risk: where each action has several possible outcomes for which the probabilities are known. (3) Uncertainty; where each action has several possible oucomes for which the probabilities are unknown.

Value:

Central to the decision theory is the concept of value, the measure of what is good or desirable about a design. Value of a product may be measured in terms of least cost, long life, least weight, be quite and fast etc. Which characteristic we must choose or coose more than one characteristic? This is very difficult and yet it is fundamental to decision-making and design in general. Value may be considered an inherent property of an engineering design. That is enginner must design a product so that it generates maximum value for the user. The amount of value will depend on the quantity assigned to the variables. Thus, a low weight may have high value, high manufacturing cost may have low value, long life may have high value, and so on. But the designer must also trade off variables. For example, he may have to sacrifice long life for low manufacturing cost. The designer must then select the best combination of engineering elements to synthesize a design, and he must juggle the quantities assigned to variables to maximize the total value potential. Engineer most of the times consider only utility values but there are other kind of values like social values, intellectual value, asthetic value, moral values, utility value, material values, technological values , basic value etc. of equal importance that should not be ignored by designer. however, it is difficult to include these nonutility values in analytical approach.

Value and Optimization Concept:

Utility and optimization are intimately linked and cannot be discussed independently. utility is really the quiantity which we maximize in optimization. Utility is dependent on such design variables as speed, weight, reliability and the like.

In design we use feasibility equations to predict performance. But if we are treating uncertainty in analytical way we must treat the variable in the feasibility equations as random. The design problems are inherently nonlinear in nature. This enormously complicates their use.

Decision maing ranks with innovation in importance in the design process. Analytical decision theory permits the engineer to model the decision process, so that he can apply the user's subjective values in an analytical way in the decision process. It tells us if a design is acceptable, that is, if it meets certain specifications with desired probability. And it helps us decide which alternate design is best; it compares configurations.

RELIABILITY should also be considered where life is considered as random.

Optimization in Engineering Design:

The primary optimization occurs when the best combonation of elements has been selected to synthesize the device. Or, on a higher level, the primary optimization occurs when the best combination of devices has been selected to synthesize a system. The use of word best immediatly requires a value judgemenr from the designer. Most of these is done intuitively. he must use logic to elect his criteia quantities, and he might use logic to decide on their order of importance. In optimization we go from primary optimization to secondary optimization to tertiary optimization.

Basic problem of application of optimization is mathematical modeling . There should be scientific way of mathematical modeling so that those who are not conversant with optimization can be helped to model the problem. Then from a large number of optimization techniques available a process of selection should be initiated for choosing the most suitable technique. This could be based on artificial intelligence or expert system. After linking the problem to a suitable technique the solution could be obtained easily. Each mathematical model may have one or more objective functions and if more than every objectives must be assigned some values. The other part of mathematical model is constraints. The formulation of contraints should also be automatic. For a novice in optimization it may not be possible to think or formulate constraints. Author has developed a method for generation of constraints based on set theory. The constraints themselves present the values and their generation based on on some knowledge based system is desirable. Author has developed an algorithm based on set theory for generation of constraints.

Total Product Life Cycle

As we approach the millenium, the rapid progress being made in advanced information technologies is prompting major changes in every facet of society, economics, the environment and manufacturing. Multimedia network and electronic commerce based on internet have already been spreading among the enterprises and consumers. Greater emhasis is also beng placed on highly individual life styles (clothing, food and housing), with the result that products are becoming more personalized.

With the advent of such progress, radical changes are also taking place in the ways of thinking with which the manufacturing sectors design their products as well as production systems. A product, hitherto assessed mainly in terms of the profit it will bring to the manufacturer, will come to be evaluated by its performance through out its life cycle based on such terms as how human friendly it is, and to what extent it helps man to coexist with natural environment while preserving the planet. Manufacturing systems will be assessed not only for their efficiency or productivity as has been done, but for its human and eco-friendlyness as well as, and this, from the viewpoint not only of manufacturers but also of consuners. of their product. It would be natural to assume that next-generation manufacturing systems incorporating such concepts as CIM and CALS should be developed based upon these criteria.

To succesfully cope with such situation, it would be critical to develop a set of basic technology to evaluate the performance of a product, or a manufacturing system, by means of a human sensory evaluation system capable of sensing their-not-so well-defined characteristics such as those which are captured by human being, when assessing their goodness by intuition. Such technology can well be taken advantage of when designing a new product, or a new production environment.

In this project, work will be undertaken to define new concept of product life-cycle, as well as to identify the role of human sensory evaluation system in conformance with definition. Further work will be carried out to establish, by taking advantage of relevent data already available, a set of technological bases upon which to develop such a system as herein proposed. Efforts will be made to extend the scope of the projects so as to finally work out a general plan for a practical model universally applicable to all possible situations.

Value Added of the International Collaboration:

Japanese industries have been maintaining a top level conditions concerning to production technology and manufacturing system, such as automization and quality control technique. recently, they begun to convert conventional lines to U-shaped lines in which producers handle multiple tasks, so that their motivation and fullfilment are enhanced.

On the otherhand, image processing technologies which are main technological field in our project have advanced in the United States and Europe, including military and medical applications. Information structures such as Internet and CALS have been researched based on the leadership of Europe and USA.

Through out the international collaboration, we expect to harmonized these advanced manufacturing methods and technologies which stored in global areas.

We would like to clear up a new product life concept presented by our project, with a global viewpoint. The aims of our project are the following:

1) Consumer-orinted product life cycle.

+Design and Develop the products in which personality and taste of consumers are reflected.

+Optimize the products based on human sensory evaluation.

2) Human friendly product life cucle

+Realize a human -machine cooperative manufacturing system which emphasize on the inherent human nature of producers.

+Realize the high-accuracyassemble and inspection processes based on the simulating a human sensory.

3) Eco-friendly product life cycle

+Coexist with earth's environment and preventing the destruction of its ecosystem.

4) Replacement or retirement of the product:

Author thinks that the total life cycle design should be based on cost minimization.

The reduction of total cost is crucial to the success of most products. Product costs are basically caused by expenses until shipment, the expenses during the product lifetime, and the expenses due to disposal and/or recycling at the end -of-life. The author considers are these costs highly nonlinear as well as stochastic in nature. Hence, an efficient mathematical tool should also be developed. For this author has used stochastic geometric programming technique and design parameters evaluated on minimization of total cost seems to work well. For a large scale decision -based design problem as proposed here some sort of decomposition approach is also important. Author tried this approach and finally recomposed it for the whole system. The algorithm seems to present a robust deign for the complete life cycle.

I will conclude it with summarizing what I consider as "Position Paper":

1) Characteristics of a designer.

2) Decision-Based design concepts.

3) Values in design

4) Total product life cycle

5) Stochastic algorithm

6) Automatic development of objective functions based on the values needed by consumers. This is almost like a multi-objective formulation.

7) Automatic generation of constraints; this is based on knowledge-based approach. Most of the users donot know how to develop constraints, hence, a comprehensive discussion in needed.

8) Which optimization technique to use? An intelligent rule-based method is to developed for proper selection of the mathematical programming technique. A through discussion on this topic is essential.

9) Stochastic cost estimation technique for adequate design is essential. The success or failure of a product designed depends on the cost of the product. Hence, a though discussion on cost estimation seems to be logical. The author seems to believe that the present day global economy needs cost estimation integrated in desgning a product. Too long this has been left in hands of accountants after the part is manufactured and total life cycle cost is most often never estimated. The author has some research work in this field and they are presented in the references.

10) Integration of recycling and costs into the product development. Very important issue which we designer always foget but very useful to consumers. A through discussion on this aspect is necessary.

References:

"Integrated Optimal Design of Manufacturing Systema", Journal of Materials Processing Technology, 1997 (accepted for publication).

"Integrated Design of Agile Manufacturing Systems", Proc. International Conference on Agile Manufacturing (ICAM'97), Feb.1997, Lafayette, LA.

"Automatic Optimal Design of Milling Cutters", Journal of Manufacturing Science and Engineering, ASME Trans., 1996 (in press)

"Mathematical Programming Formulation and Optimal Design of Robot Control System", Applied Mathematical Modeling, 1996, Vol. 20, November, pp. 853-865.

"Integrated Computer Aided Optimal Design and Finite Element Analysis of a Plain Milling

Cutter" Applied Mathematical Modeling 19(6) 343-353, 1995

"Optimal Design of a Robot Control System Through Nonlinear Programming," International Journal of SYSTEMS AUTOMATION: RESEARCH & APPLICATIONS (SARA) , 2, 254-275 (1992)

"Computer Aided Formulation and Optimal Control of an Integrated Production Systems of Competitive Products," International Journal of Production Planning and Control, 1994, Vol. 5, No. 1, 47-58.

"Optimal Image correlation in Experimental Mechanics", Optical Engineering, April 94, Vol.33, No.4, pp.1099-1105.

"Computer-Aided Multi-Objective Optimal Design and Finite Element Analysis of Cutting Tool,", Proc. ASME Manufacturing Internation 1990, Atlanta, GA, March 1990.

Computer Aided Optimal Design and Finite Element Analysis of Plain Milling Cutter," Computers in Engineering, Vol.2, pp. 511-521, ASME 1989.

"Computer Aided Optimization and Design for Manufacturing," Computers in Engineering, Vol.1, pp. 105-112, ASME 1987.

"Engineering Optimization and Expert System," Proc. 1985 ASME Computers in Engineering Conference, Vol. 1, pp. 97-101.

"A Set-theoretic Automatic Selection of Cutting Tools in Manufacturing Systems", Journal of Materials Processing Technology, 65 (1997) 99-106.

"Automatic Optimal Design of Milling Cutters", Journal of Manufacturing Science and Engineering, ASME Trans., 1996 (in press)

"Probablistic Cost Estimation in Advance of Production in a Computerized Manufacturing System Through Stochastic Geometric Programming", Computers & Industrial Engineering, Vol. 30, No.4, pp. 809-821, 1996.

"Computer Aided Optimal Turning of Stepped Shaft", Journal of Materials Processing Technology, 57 (1996) 38-47.

"Integrated Computer Aided Optimal Design and Finite Element Analysis of a Plain Milling Cutter" Applied Mathematical Modeling 19(6) 343-353, 1995

"Manufacturing Cost Estimation in Computer Integrated Manufacturing System," Flexible Automation and Intelligent Manufacturing (FAIM '95), Stuttgart, Germany, June 1995.

"Manufacturing System Design, Analysis and Control," Applications of Management Science, Vol.9, 1996. Kenneth D. Lawrence & Gary R. Reeves Edited).

"Stochastic Mathematical Modeling and Manufacturing Cost Estimation in Uncertain Industrial Environment," International Journal of Production Research, 1992, Vol. 12, 2755-2772.

"Unified Theory of Automation in Process Planning Through Multi-objectives

Under Real World Constraints," Applied Mathematical Modeling, Vol.16, pp. 58-69, 1992.

Expert System in Manufacturing Planning and Multi-objective Optimum Process Planning of Milling Operation- a Comparison," Proc. 1991 ASME Computers in Engineering , July 1991.

"Flexible Manufacturing Systems Handbook," Academic Press, July 1991.

"Integration of CAD to CAM Through Optimum Process Planning and Expert System Approach," Computers in Engineering, Vol.1, ASME 1991, pp. 43-51.

"A Discrete Data Base Multiple Objective Optimization of Milling Operation Through Geometric Programming," ASME Trans., Journal of Engineering for Industry, Vol.112, No.4, 1990.

"Optimum Design of Manufacturing Systems Through Set-Theoretic Approach," ASME Manufacturing International 1990, Atlanta, GA, March 1990.

"Multi-Objective Theory of Automation in Process Planning," ASME Manufacturing International, pp. 191-201, 1988.

"Optimum Process Planning (OPIPP) and Expert System Process Planning (ESYPP); A Key to CAD/CAM Integration," Computers in Engineering, Vol. 2, pp. 451-458., ASME 1988.

"Computerized Process Planning of Grinding for Multiple Objectives Through Geometric Programming," Computers in Engineering, Vol. 3, pp. 295-302, 1987.

"Optimizing the number of tools and cutting parameters in multi-tool turning for multiple objectives through geometric programming," Applied Mathematical Modeling, 1986, Vol.10, pp.163-169, June.

"Flexible Manufacturing Facilities Selection in Uncertain Situation," Computers in Engineering, Vol.2, pp.203-206, 1986.

"Engineering Optimization and Expert Systems," Computers in Engineering, Vol. 1, pp. 97-101, 1985.