2
DECISION MAKING IN ENGINEERING DESIGN
The role of decision making in an engineering design context can be defined in several ways. As shown in Figure 2–1, the decision process is influenced by sets of conditions or contexts.
The business context represents the long-term view of the engineering company and is largely in the control of the company. The environmental context, such as the state of the economy, is not controlled by the company and must be considered a variable. The input context, such as the completeness of and variation in requirements and constraints, is established by the customers as is the output context, such as state of readiness to implement decisions, risks, and qualifiers.
Closing with the customer is an iterative process reconciling the customer’s needs with the developer’s design capabilities and requiring collaboration and experience with the product. In the real world, decisions made by the experts can be delayed and overturned by higher-level management based on poorly defined or unstated environmental issues.
In today’s engineering environment routine decisions can involve geographically dispersed teams working under challenging cost and timing constraints. Under these conditions, the quality of most decisions can be improved through the application of computer-based tools. These tools can be put in the following categories: knowledge-based engineering, workflow, and collaboration.
Knowledge-based engineering tools provide computational representations of engineering design rules, allowing engineers to execute modeling, analysis, and optimization far more quickly while
staying within design practice constraints. This frees the engineer from the routine portions of analysis and allows more time to trade design options, thereby improving design decisions.
Work process management tools can help manage the execution and coordination of tasks via the Internet by routing and tracking work throughout the design process. These tools manage both schedule execution and process compliance (such as configuration control) globally.
Collaboration tools, including Internet-based conferencing and graphics sharing, help the day-to-day working relationship between engineers at distributed locations. This ongoing real-time collaboration results in timely and coordinated design decisions.
Figure 2–2 shows the character of decisions affected by two elements of the business context. On one axis is the duration of the decision and on the other is the criticality of that duration for the product. A decision in this context is a choice made by the design engineer for a particular solution for the problem at hand. Decisions with long-term impact often are irreversible after implementation; therefore the decision maker must seriously analyze the decision. A large number of short-term incremental decisions can, however, be made relatively risk free. All decisions can be plotted on the context chart and fall into one of the relevant subgroups. Increasing investment of scarce time and other resources in the decision process is appropriate for the decisions that are critical or irreversible.
Correctly assessing the context for making a decision is important because it dictates the level of effort and type of tools applied to the decision process. The most critical decisions often employ multiple tools, preferably logically sound and internally consistent in all circumstances. Framing addresses how to pose the decisions, and is described more fully in Chapter 3.
Table 2–1 Framing a Decision in the Relevant Context
|
Routine Decision |
Make-or-Break Decision |
||
Context |
• |
Small impact |
• |
High impact |
• |
Reversible |
• |
Irreversible |
|
• |
Short term |
• |
Long term |
|
• |
Data available |
• |
Safety or product liability critical |
|
• |
Standard decision process |
• |
No well-defined decision process |
|
Decision Processes and Tools |
• |
Use standard or automated decision process |
• |
Use a variety of decision processes |
• |
Supporting data available |
• |
Generate supporting data |
|
• |
Use simple decision tools |
• |
All functions involved |
|
• |
Small team or individual |
• |
Large team including management |
|
• |
Low-level reviews |
|
Decisions that make or break the business (Table 2–1) are often laden with trade-offs, which are usually complex in nature. Further, preferences for these attributes typically differ across stakeholders, such as customers, operators, and manufacturers. Framing and resolving these trade-offs are time pressured with much potentially relevant information to be considered. These trade-offs are also subject to many uncertainties regarding customer buying preferences, user abilities and preferences, technology maturity and availability, and competitive advantages of possible functions and features. These trade-offs usually cut across disciplinary boundaries in terms of balancing weight, power, speed, cost, and economy of use. In some cases these trade-offs are resolved by fixing design requirements, which makes design more tractable but increases the chances of noncompetitive solutions because of a restricted ability to trade off design options. Most contemporary design methodologies avoid premature freezing of requirements.
The complexities just portrayed result in multi-disciplinary teams for most design efforts. The notion of such teams once implied multiple disciplines, such as mechanical and electrical engineers, working together. More recently, however, multi-disciplinary has come to mean engineers, industrial designers, marketing and sales professionals, and finance experts working together. This enables richer, more comprehensive trade-offs across form, functions, features, and price. This challenges many design decision tools because of the bias and limitations of individual disciplines. Richness in this context refers to the wealth of relevant information and the many players in the decision making process. Because conceptual and mathematical representations of different disciplines do not easily mesh, it is difficult to reach common analytical solutions. Either each discipline tends to sub-optimize their piece of the problem or, more likely, decisions are made more subjectively through negotiation rather than calculation.
Economic attributes ranging from financial metrics to consumer utility models cut across all disciplines, especially when supported by the methods and models of probability and statistics. These models and methods allow exploration of the intersections of market-driven preference spaces and technology-driven physical spaces.
Economic attributes drive actual design decision making, regardless of the extent to which the methods and tools include such attributes. Similarly, uncertainty and risks are pervasive and must be addressed. These overarching issues must be considered regardless of the decision making tools used. These cross-cutting approaches provide the context for design engineers to synthesize and analyze ways of providing desired functions and features within economic constraints, as well as quality,
reliability, and maintainability considerations. Analyses of signal flow, stress characteristics, and control stability are in this way integrated into the overall design context.
Traditional design engineering is still pursued, but it is less and less isolated by trade-offs and optimization within a discipline-limited set of purely physical variables. This is nowhere more evident than in the linkages of design variables to economic considerations. Representations of interactions of product and process variables on costs have become central to product realization in many domains. Multi-attribute, multi-stakeholder design contexts, laced with uncertainties and rich in information, are the norm. The framing of critical design decisions across contributing disciplines is central to success in such contexts.