Thomas Friedman, the author of Hot, Flat, and Crowded, has invited readers to contribute ideas for a final chapter for the second version of the book. He wants to hear our thoughts on how we might ‘grow people’s living standards in a more sustainable and regenerative way’. (If you haven’t yet read HFC, I highly recommend it.) Here is my response to Friedman’s invitation:
In Hot, Flat, and Crowded you discuss the importance of ’smarter’ design; by changing how things are built, how they work, and are retired, we can reduce energy consumption and environmental impact dramatically, as well as improve quality of life and national security. I believe better design is at the core of a green revolution, and we need increased efforts to help others solidify mental links between design improvements and a vision for a sustainable future. In addition to helping citizens deepen their appreciation for the role of design, we must address this issue on two other fronts: public policy and engineering expertise. We need the right policy and incentives to set the stage for a transition to sustainability, as well as the technical expertise to implement the transition rapidly. I would like to address the latter issue.
To realize a green revolution, we can’t settle for products that are ‘good enough’, or green technology that evolves slowly. Instead, we must seek to develop the very best, most efficient designs, and do so quickly. Instead of taking small steps each year with slightly more efficient cars, slightly better wind turbines, let’s make giant leaps! We need the backing of citizens, the support of policy makers, and boldness from engineers and engineering educators to advance our ability to create sustainable systems and products. Researchers have developed impressive new engineering design methods the last few decades that can help us create products and systems that use less energy and other resources, while making leaps forward in performance. Some of these methods are mature and proven, but unfortunately are not yet used widely by engineers. First, let’s have a look at the conventional design process.
Suppose we were designing a car to be very energy efficient, but still performs well at a reasonable cost. Using a conventional design process, engineers would generate design ideas, test these candidate designs, propose new designs, and iterate until they converge on a design that meets (or comes close to) design targets. In the past, engineers relied heavily on expensive physical prototypes for testing. More firms now use computer models that predict how something will perform without having to build it. While this saves time and money, design refinements often are still made by engineers based on test results, experience, and expertise. Managing all these often conflicting design decisions is often overwhelming, particularly as products evolve and become more complicated; engineers stop when they find a design that meets basic requirements, instead of pursuing the best possible, or optimal, design.
One prominent method developed by researchers is design optimization. Other readers have also described optimization as an important solution; I hope to strengthen this position and clarify the link between optimization and engineering design. When using design optimization, engineers work to minimize or maximize some important aspect of a product, in addition to seeking to meet design requirements. In the car example, we might seek to maximize fuel economy, while meeting acceleration, handling, comfort, cost, safety, and other constraints. Framing a design problem in this way allows engineers to use computer models and powerful optimization algorithms together to help generate the best possible design. In this process design candidates to be tested are chosen analytically using mathematical techniques, reducing the number of tests and time to market. It can help engineers learn what is really achievable, opening our eyes to new possibilities. Design optimization also accelerates design evolution by enabling engineers to make more substantial design changes between product generations, instead of just small perturbations of the last version (as is usually the case now).
The design optimization approach is actually a pretty natural fit for how engineers already go about designing things; using formal design optimization is an enhancement that produces better results in less time, and leverages investments many firms have already made in computer modeling. It’s not a push-button solution; it automates some aspects of design, but requires engineering expertise and experience to implement successfully. (In the parlance of The World is Flat, design optimization is a high-level, ‘icing’ activity). Awareness is perhaps the biggest hindrance to the adoption of design optimization. It needs to be taught in undergraduate (not just graduate) engineering courses, as well as in industry training programs.
In summary, design engineers make a lot of important decisions that have tremendous impact on our world. Moving beyond status quo design processes can help engineers deliver sustainable products and systems while improving living standards; these changes in engineering design are essential to a successful green revolution. Right now there is a lot of low-hanging fruit; there are many opportunities to improve our world through better design. Design optimization can help us put new technology into production faster, as well as refine systems that use existing technology. This can help us bring energy efficient designs into production more quickly, and accelerate the transition to renewable energy systems. We have the technical tools, but we need the societal impetus to put them to broad use.
James T. Allison, Ph.D.