Decision-Based Design: Where Should We Go from Here?

Janet K. Allen
janet.allen@me.gatech.edu

Farrokh Mistree
farrokh.mistree@me.gatech.edu

Systems Realization Laboratory
Georgia Institute of Technology
Atlanta, Georgia 30332-0405

In the engineering community, there is an emerging understanding that a framework for engineering design is decision-making and that decisions are the fundamental construct in engineering design. Decision-Based Design, DBD, is developed from this perspective. In order to develop DBD as a foundation for future progress in design, we need to develop a shared understanding of the definitions of the fundamental terms, specifically, design, decisions, and, further, Decision-Based Design. There are several definitions for these terms in the literature and also in standard English dictionaries; the importance of developing a common vocabulary is well articulated by Dixon. As a working group it is essential for us to determine which flavor of definition we will use and how we will build upon it. We suggest that we agree upon definitions of the following terms so that we can develop a shared understanding of these terms, some definitions are given here

Designing:
Designing is a process of converting information that characterizes the needs and requirements for a product into knowledge about a product.

Decision:
"...we define a decision as an irrevocable allocation of resources." "There are two important characteristics of a decision: A decision is made at an instant in time. A decision must be made based on the information available at the time it is made."

Decision-Based Design:
a term coined to emphasize a different perspective from which to develop methods for design. The principal role of a designer, in Decision-Based Design, is to make decisions. This seemingly limited role is useful in providing a starting point for developing design methods based on paradigms that spring from the perspective of decisions made by designers (who may use computers) as opposed to design that is assisted by the use of computers, optimization methods (computer-aided design optimization) or methods that evolve from specific analysis tools such as finite element analysis. Decisions help bridge the gap between an idea and reality. In general, decisions are characterized by information from many sources (and disciplines) and may have wide ranging repercussions. In Decision-Based Design, decisions serve as markers to identify the progression of a design from initiation to implementation to termination. In Decision-Based Design they represent a unit of communication; one that has both domain-dependent and domain-independent features.

The following is an assessment of the state of engineering design, Decision-Based Design and our role in this workshop. It is intended as a framework to stimulate discussion to help clarify our position: comments, additions, suggestions and disagreements are welcome.

Assumptions in Decision-Based Design

Inherent in Decision-Based Design, there are at least two assumptions:

Decisions are fundamental design constructs.

Although design is a science of the artificial and hence is created to support humans, that human abilities and the type of design possible in the industrial environment are relatively static. Given that there is great change in the industrial environment as far its technical capabilities, we assume that in the foreseeable future, humans will continue to guide the design process.

From our perspective, that of the Decision Support Problem Technique, we add the further axiom that there are two major types of decision, the selection decision (either-or) and the compromise (trade-off) decision. It is further worth noting that selection may be thought of as a particular form of compromise. Further we anchor our approach, the Decision Support Problem Technique in the paradigm of Living Systems Theory.

Group goals for the Decision-Based Design Workshop (as enunciated in the workshop announcement)

The goal of this open workshop is to engage design theory researchers in a dialogue to establish a rigorous and common foundation for research and educational endeavors.

to establish the role of decision-based design in mechanical engineering, engineering in general, and in a modern social environment by debating issues such as "Is decision making the fundamental construct in design?" and "Is decision-based design founded in behavioral studies, social sciences, or in physical sciences and technology?"

to provide a foundation (approaches, problems, references, textbooks) that could both illustrate design as decision making and be used in classrooms to help students gain experience in decision analysis.

to define what "design" should and should not be from a decision making perspective.

to establish a relationship between the theories developed in other science domains such as physics, mathematics, information and management science with the theories developed in design engineering methodologies.

From this perspective, there seem to be two distinct activities involved in our group goals: (i) creating scholarship and (ii) developing consensus, and the focus of our activities is within academia, rather than in industry. It is worthwhile noting that the majority of the proposed activities and effort for this workshop appear to address the former, rather than the later issue.

Scholarship in Decision-Based Design:

Can Decision-Based Design form the foundation for research areas that will have the greatest impact on engineering design over the next ten years

The NSF Strategic Planning Workshop help in April of 1996 identified the following research areas of great importance in the future:

In addition, it seems appropriate to understand how existing work fits in with DBD, e.g., DBD/Pahl & Beitz, DBD/Boothroyd & Dewhurst. Is there a body of existing engineering design that is not DBD? In the interests of building consensus, it is worth considering how each of these fits in with Decision-Based Design.

What then are appropriate areas relating to scholarship in Decision-Based Design?

As a minimum it seems appropriate to consider:

Various methods which have been proposed for decision support and structuring decisions

Benchmark cases and software

Efficiency in decision making, keeping track of decisions and integrating groups, computational methods, implementation platforms

Effectiveness in decision making

Levels of information needed at different times

Evaluation, both final product and ongoing during process

Distributed decisions - modularity of decision making units versus distributed decision making and integration issues.

Access to and use of information

Design evolution - how can you tell where you are in a large scale design process - issues: iteration, complexity, starting point (concept exploration), customer/market demands (problem formulation)

Issues of representation and communication

Dealing with uncertainty.

Suitable areas for scholarship in design pedagogy:

What should an undergraduate engineer know/be able to do in design? A graduate engineer?

What are suitable representation schemes for decisions, and for the design process overall. Also for function/behavior/structure of systems.

What is the most effective way to assess studentsí design knowledge?

What information management tools are useful to a student

Efficiency in education (more with less time, faculty, resources)

What educational materials are useful: software? textbooks? problems? sample projects

What kinds of scaffolding are necessary, when in design process

how to fade the scaffolding. (Scaffolding relates to instructions or advice that the student is given just in time and the assistance in accomplishing mundane tasks which allow him/her to focus on major learning issues.)

Deliverables - a Question

It is often hypothesized that we are moving toward the science of design. In general are our goals as a group to extend/enhance the extent to which design becomes a science? If so are we looking for hypotheses which then become testable theories on which we can build an intellectual structure?

Another Question :

The focus of this workshop is basically academic design. What do we perceive the role of industry to be?

Issues Relating to Building Consensus.

Here the situation is not as clear, it seems that we, at least, have not fully understood the reason that DBD has not been embraced by the academic community. Some possibilities follow:

Barriers to consensus:

Competing paradigms? Do we need to demonstrate that DBD is better than other design paradigms?

Or perhaps the belief is that there can be no paradigms for design? The idea that design is somehow essentially creative and that either someone can do it or cannot do it ... and that it happens in a flash of inspiration.

The need to perform unique research

The lack of a clearly articulated position.

Available materials: textbooks, etc.

Education of faculty members

Academic inertia

The Million Dollar Questions: How do the activities we will undertake tie in with the goals of improving scholarship in Decision Based Design and building consensus? Do we have the balance right? What are the differences between what we will do as a group and our individual research?