ILP Institute InsiderMarch 13, 2017
Clearing Away Complexity
Edward Crawley brings system architecture tools to tough technical decisions and rethinks how to educate the engineers who make those decisions.
When engineers start to model a truly complex new engineering project such as a jet aircraft or a refinery or an automobile or a Mars mission, they have two basic choices. Most commonly, they reiterate on previous design solutions that have proven reasonably successful.
Alternatively, they can start with a clean slate—modeling every critical early design decision point, examining all the resulting combinations, analyzing the payoffs and risks and costs of each combination, and identifying the most promising approaches to pursue.
That’s the “system architecture” method pursued by Ed Crawley, MIT Ford Professor of Engineering and professor of aeronautics and astronautics.
“The need today to put more features, capability and performance into products, and to network all of our systems together to build systems of systems, is creating this enormous issue of how to reason about complex systems,” says Crawley.
He and his colleagues at MIT’s System Architecture Lab tackle this issue with software tools that offer quantitative frameworks for decisions on complex technical systems and their associated economic and stakeholder constraints.
Crawley, who has held a number of leadership positions in engineering education, also focuses on ways to strengthen teaching for MIT engineering students. As one key to that effort, he seeks advice from Industrial Liaison Program (ILP) members about their needs for well-trained professional work forces.
Widening the scope for optimization
Deeply complex products such as commercial aircraft and operations such as Mars missions are driven by key decisions made early in each program's lifecycle, Crawley notes. System architecture aims to offer a framework that captures all of the possible decisions, rather than working from previous decisions.
“The System Architecture Lab is looking at techniques, processes and tools to support decisions about the crafting of very complex systems of all types—electronic, informational, mechanical or biological,” Crawley says. “We can apply these tools to all types of problems, most importantly to problems that transcend traditional disciplines.”
“We’ve developed techniques and design environments in which this type of high-level decision making can be supported,” he says. “You have to recast the problem into the decisions that the decision makers will eventually make. When you do it this way, it becomes much simpler for the decision-makers, to be able to understand and reason through the complexity: If you choose this, you get this result, with its associated costs and benefits.”
System architecture builds computational models of the network of events, and then models of the pieces of form that will perform these sequences of events, Crawley explains. Next, the software combines these in all possible viable combinations into an integrated system, which might hold millions or tens of millions of examples. Then the decision makers can evaluate the performance of each of those tens of millions of examples, by criteria such as cost or the probability of mission success.
This approach has proven useful in early planning for human travel to Mars, which he has studied extensively with the National Aeronautics and Space Administration. Crawley and his colleagues also have made contributions to optimizing commercial jets, earth observation from space, space communications, offshore oil exploration, refinery design and development, and other complex applications.
The control of the new machines
Crawley has long focused on ways to improve engineering education—not just as MIT professor but as head of the Department of Aerospace and Astronautics, director of the Bernard M. Gordon MIT Engineering Leadership Program, and most recently as the founding president of Skolkovo Institute of Science and Technology in Moscow.
He’s now working with other faculty on the Institute’s NEET (new engineering education transformation) initiative, which seeks to bring “a continuous process of improvement” to engineering education at MIT.
“One of the main themes of this deep look at our education is the idea of the new machines: What will our students build in the middle of their careers, in the middle of this century?” he says.
Crawley sees three main trends: The new machines will be increasingly complex, they will integrate interdisciplinary technologies and they will need to use sustainable energy sources and materials.
So how can MIT best prepare students to build these machines?
As the NEET initiative studies the role of a university like MIT, “one factor is that the generation of new scientific and technical information is accelerating,” he says. “The idea that you can send somebody to college and they get a lot of just-in-case information is an idea of the past. This generation of students can teach themselves almost any facts they need to learn on the Internet in a manner of hours.”
“We’re coming to the view that the value is teaching students how to think, not in a general way but in a set of very specific ways,” says Crawley. “Thinking about analytics, computation and experimentation; thinking critically and creatively; thinking about systems; and thinking about humanism, values and working with others.”
Additionally, engineering schools must prepare students for all the roles they might eventually fill. “Our graduates will work on a spectrum from being discoverers (that is to say researchers) to being makers (that is to say engineers),” he notes “We should prepare them all along that spectrum.”
Working extensively with industry over his career, Crawley also, like many MIT faculty members, is a serial entrepreneur. He has cofounded four startups that move technologies from aerospace into other commercial markets: biomedicine, manufacturing, online services and construction.
As he and his NEET colleagues try to figure the road ahead, “we very much want this education to be aligned with the needs of industry, both the industry of today and the industry of tomorrow,” Crawley emphasizes. “We’re engaged with ILP members in surveying their views and needs, as one way to get the voice of industry to help guide our solutions.”
Education Partner: Professional Education
Join Prof. Crawley’s short course Systems Engineering and Architecture: Principles, Models, Tools and Applications, July 17-21 at MIT Professional Education. Learn the foundations of SA as a series of decisions that frame form to function mapping, as well a number of architecture representations and methodologies. Learn more.
More ILP News
- Human-Machine Interaction and Innovation in Healthcare March 20, 2017
- The Intersection of Sports, Technology, and Beyond March 8, 2017
- Make Your Business Culture Move February 13, 2017
- Getting Inside Osteoarthritis February 6, 2017
- Taking Flight with Numerical Models January 23, 2017
- Optimizing the New Networks January 17, 2017