Welcome, Guest  |  Login    Register

Conquering Complexity: Learn From Your Mistakes Faster

Posted By Jeff Moad, April 17, 2012 at 7:49 AM, in Category: The Innovation Enterprise

I recently finished reading a sprawling, fascinating novel called Shantaram that follows the exploits of an Australian prison escapee who finds his way to Mumbai, India, where he becomes, at various times, an untrained slum doctor, a heroin addict, a gangster, and a military mercenary. At 930 pages, it’s like Homer’s Odyssey on the subcontinent.

One of the work’s more interesting characters is an avuncular though ruthless mob boss who has his own theory about how the universe works and how one can distinguish acts of evil from acts of good.

"The universe began about 15 billion years ago, in almost absolute simplicity, and it's been getting more and more complex ever since,” says the mob boss. “This movement from the simple to the complex is built into the web and weave of the universe, and it's called the tendency toward complexity. We're the products of this complexification, and so are the birds, and the bees, and the trees, and the stars, and even the galaxies of stars....The final or ultimate complexity—the place where all this complexity is going—is what, or who, we might call God. And anything that promotes, enhances, or accelerates this movement toward God is good. Anything that inhibits, impedes, or prevents it is evil."

While I’m not sure I agree with the mob boss’ definition of God (and while I don’t really approve of his use of a word like “complexification”), I do recognize what he and others describe as the inexorable tendency toward complexity. Just think about some of the products and tools most of us touch—and some of us make—every day. Cars, commercial aircraft, medical equipment, etc. All have become more and more complex as they’ve taken on new capabilities and as most have incorporated higher levels of intelligence, thanks to the availability of cheap electronics and the ubiquity of the Internet.

And as the complexity of these types of products has increased, the way they are designed, built, and tested has changed, not always for the better. For one thing, it’s becoming harder for small companies and individuals to play a meaningful role in the design and innovation of very large, complex products. Instead, the very specialized capabilities needed to design, simulate, test, and redesign these complex systems are being consolidated in fewer and fewer companies such as large defense contractors.

And as these systems and products become more complex, it becomes harder for even these mega-firms to anticipate all of the unintended consequences of design decisions. That’s why it can take a decade or more for, say, a complex commercial airplane or defense system to go into full production.

“We’ve seen a paradigm shift where, as you go to more complex mechanical systems, there are only small numbers of groups that can approach these problems today, who can bid on and make these systems,” says Joe Salvo, manager of the business integration technologies laboratory at GE Global Research, with whom I spoke last week.

“At the same time,” Salvo says, “as systems get so complex, you can no longer use traditional methods to determine how they will behave until you build them. Then you have to decompose the design to address unexpected behaviors. So, the cycles are extended, and the development time and cost of development go well beyond where you want them to be.”

Salvo is involved in an experiment designed to apply the notion of crowdsourcing to accelerate the design of complex industrial products. Called (Vehicleforge.mil), the effort is being undertaken by the military’s Defense Advanced Research Projects Agency (DARPA), GE, and several universities. The idea is to build and make available a set of design technologies and a project methodology that will allow even small groups and individuals from around the world to make meaningful contributions to the development of complex industrial products—and to benefit financially from those contributions.

The project is focusing first on subsystems to be used in a military amphibious vehicle. The project’s first phase early next year will award $1 million for the winning drive-train system design. Another $1 million will be handed out later next year for the winning chassis design. And in 2014, a $2 million prize will be awarded for the best overall vehicle design.

GE is working with MIT, Georgia Tech, and other universities to develop a Web-based platform that will include collaboration and simulation tools as well as pertinent content.

And DARPA is putting together the ground rules on how these crowdsourcing collaborations would work. How, for example, will winning subsystem designs be tested and integrated into an overall vehicle design? And how will winning designers in different parts of the world physically test the overall designs and integration before production begins?

Even in fields such as software development, where crowdsourcing has gained more traction, critics point out that integration of subsystems into a coherent overall system design can pose a hefty challenge. Some question whether crowdsourcing can translate to the design and testing of large, complex industrial systems where integration can be a greater challenge.

Salvo acknowledges that Vehicleforge.mil has attracted it share of doubters. “Because this is such a bold idea, there will always be critics and naysayers who want to emphasize the most extreme cases to make their case that this will be too difficult,” he says.

But, he suggests, Vehicleforge.mil would make a valuable contribution even if the project doesn’t lead to the selection of a brilliant, winning amphibious vehicle design submitted by some genius working in his garage. Just as likely, he says, the project will generate lots of great ideas from individuals and small groups who can spot and point out design problems well before a more insular group working within a traditional defense contractor.

“A lot of people enjoy trying to take things apart and trying to break them,” Salvo notes. “They usually aren’t viewed as a positive force in the creativity process, but that’s wrong. These people will be able to help us fix the designs so we don’t put all our effort into what we think is right, only to have to change it later.”

In other words, Salvo says, the ultimate benefit of crowdsourcing for complex industrial products might be that it will allow design mistakes to be found and fixed much faster.

Written by Jeff Moad

Jeff Moad is Research Director and Executive Editor with the Manufacturing Leadership Community. He also directs the Manufacturing Leadership Awards Program. Follow our LinkedIn Groups: Manufacturing Leadership Council and Manufacturing Leadership Summit

Add Pingback


You must be logged in to leave a reply. Login »