I’m up in Boston (hence the slow posting), home of that legendary creature, conductorus bostoniana — i.e., the “Boston driver” — a topic I’ll return to in another post. But just to note there’s a great write-up in the latest issue of Wired, by Josh McHugh (article here or after the jump). In it is discussed briefly a topic I’ll also want to return to in a later post: comparing internet traffic to vehicular traffic.

Tom Vanderbilt’s Why We Drive the Way We Do Unlocks How to Unclog Traffic
By Josh McHugh
Driving down a New Jersey highway three years ago, Tom Vanderbilt decided to stop being a goody-goody. He fought the urge to merge at the first indication that his lane was ending and rode it right to the pinch point, wedging his way in front of a furious driver at the last second. Racked with moral misgivings, he eventually looked into the science of merging and discovered salvation in high math, which proves he made the right choice — and not just for his own time-saving benefit, but for humankind (or at least commuter-kind — the seemingly selfish strategy keeps traffic moving faster for all). “It doesn’t have to be an ethics problem,” Vanderbilt says. “It’s really a system-optimization issue.”

That’s when he decided to write Traffic: Why We Drive the Way We Do (And What It Says About Us). As part of his research, Vanderbilt set up Google Alerts to notify him about traffic-related news. “Half were about road traffic, and half were about Internet traffic,” he says. Unfortunately, drivers have a major disadvantage relative to data packets flowing across the Web: Humans think too much. Packets go where they’re told rather than relying on the scraps of incomplete intelligence and “superstition,” as Vanderbilt calls it, that humans use when choosing how to get from point A to point B.

Drivers make shortsighted decisions based on limited information — a combination of what they can see and traffic reports that, even at their most sophisticated, are an average of 3.7 minutes old. At 60 mph, that’s a 4-mile blind spot. “The fundamental problem,” Vanderbilt says, “is that you’ve got drivers who make user-optimal rather than system-optimal decisions” — a classic case of Nash equilibrium, in which each participant, based on what they believe to be others’ strategies, sees no benefit in changing their own.

Those who seek a more efficient traffic solution use not only network topology and queuing theory but psychology and game theory, too. A typical puzzle: Waiting for an on-ramp metering light — a mild and remarkably effective congestion-control measure — has been proven to rankle drivers more than merging directly into a traffic jam. “What bothers people is that they can see traffic flowing smoothly,” Vanderbilt says. “So they think, ‘Why should I wait?’ They tend not to accept that the traffic is flowing smoothly precisely because of the metering light.”

What about faster, better traffic info? One new technology, Dash Navigation’s GPS-based social networking system, may be a step toward dynamic traffic routing, but only for those who have Dash’s device, and maybe only temporarily. Suppose Dash were to become the hit its backers — including VC firms Sequoia Capital and Kleiner Perkins Caufield & Byers — hope it will. As soon as drivers have all the information about which routes are congested, they’ll divert to others that are clear. But if enough people do this at roughly the same time, the clear routes become jammed. Vanderbilt laments this as the inevitable “death of the shortcut.”

The obvious answer, then, is to make the road network as efficient as the information superhighway. Make the packets (cars) dumb and able to take marching orders from traffic routing nodes. The obvious problem with that: No self-respecting, freedom-loving American would stand for it.

This entry was posted on Tuesday, July 22nd, 2008 at 5:45 am and is filed under Book News, Congestion, Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.