A while back, the Christain Science Monitor ran a fascinating article on one of the most counterintuitive subjects in transportation policy: the so-called Braess Paradox.  Stated simply, mathematician Dietrich Braess proved the unthinkable:  sometimes, building a new road—even a high-speed one—can slow down traffic.  (Here’s the Wikipedia page on the issue, if you’re curious. And note that Braess’s discovery can apply to computer networks as well as roads.)

This is not actually a a paradox in the strictest sense.  It’s just one of those things that sounds completely implausible, but is nonetheless completely true.  The converse is just as true: sometimes, you can speed up traffic by closing a road.  The city of Seoul, Korea, for example, identified a high-capacity highway that had this strange property, and was actually gumming up traffic.  The city pulled down the highway and—almost like magic—traffic eased a bit, even as the urban environment improved.

According to researchers, other cities could benefit from the same approach.  The CSM article highlighted a brilliantly titled study—“The Price of Anarchy in Transportation Networks“—that discusses some real-world examples of cities that might benefit from following Seoul’s example.  Transportation geeks should take a few moments to check it out.

  • From the abstract:

    Society…has to pay a price of anarchy for the lack of coordination among its members. Here we assess this price of anarchy by analyzing the travel times in road networks of several major cities. Our simulation shows that uncoordinated drivers possibly waste a considerable amount of their travel time. Counterintuitively, simply blocking certain streets can partially improve the traffic conditions.

    The authors define the “price of anarchy” as the difference between the ‘social optimum’ (the best possible performance of the road system, if everyone cooperated) and the ‘Nash equilibrium’ (real-world perfermance, in which everyone acts in their own interest with no real regard for the common good).  It’s essentially a quantification of a market failure, in which individually rational decisions lead to ridiculously irrational outcomes.

    The authors used some nifty but not-too-complicated methods to calculate the “price of anarchy” in the road networks of London, Manhattan, and Boston.  London - Braess paradox, Youn, Jeong and GastnerAnd they found that anarchy can increase travel times by as much as thirty percent.  Apparently, when people crowd into the fastest routes, they can slow down travel for everyone.

    Just as importantly, the authors identify specific roads that, if eliminated, would reduce the “price of anarchy” and improve travel times. In the map of London to the right, the dotted black lines are the offending routes:  close them to cars, and traffic would speed up!  The red lines are the vital ones: close those lines, and traffic gets significantly worse.  Blue streets—the large majority—could be severed with no significant change in travel times.  Most curious to me is the preponderance of blue routes.  Just think: a careful mathematical traffic analysis finds that the bulk of urban arterials are more or less irrelevant to overall travel time.

    The authors also analyzed Manhattan and Boston, with strikingly similar results.  I’d love for them to take a look at the road networks in the Northwest.  Perhaps we’d learn something surprising—a few streets that would be better off turned into parks or bike lanes.  At a minimum, the exercise might help debunking of the idea that the only good way to make traffic faster is build bigger, wider roads.