Ending Traffic Jams

By Eric DuBois

My post today is somewhat motivated by Sam’s comment last week about his brother avoiding stoplights even if it did not improve his overall drive time.  It reminded me of an interview I heard on the radio while hiking a couple years back.  The argument was that we could alleviate many types of traffic jams by simply moving slower and more constant rate.  There is actually quite a following for this theory and it has proven true so far when tested experimentally.

This works on two levels.  On the more objective traffic engineering level, it smooths out the ‘traffic waves’ where cars slow down in a given area (possibly to a stop) and then speed up outside of that area.  Essentially, stop and go traffic is a series of traffic waves.  This is obviously more gas efficient and safer, and better yet, it actually increases total throughput in that area.  More directly to queuing psychology, it means that rather than constantly having to stop, the driver is constantly feeling a measure of progress, which improves satisfaction with waiting in line.

A similar theory holds when traffic is trying to merge, say at an off-ramp.  If we leave room and allow drivers to merge with us, then we can all maintain a relatively constant, if slow speed.  Whereas by cramming together, the driver must force their way in and that can stop traffic entirely creating a new traffic wave

The big complaint with this is that it means allowing a gap to open between you and the driver ahead.  You need a large enough gap that in the time it takes you to reach the next car, they are moving again.  Now you may say (and I certainly have), that it is not fair to those of us waiting in line to have this new person drive past us, making us look like fools, and cut in line up ahead.  However, that extra car would only add a few seconds to your total commute and by choosing to spite them, you are making the situation noticeably worse for everyone following after.

In essence, we are focused on what queuing theory calls a skip, that is, they are skipping ahead of us in line and a slip- we are slipping further back in line.  On a very basic psychological level, this doesn’t seem fair and annoys most of us to no end (some more than others) by its violation of the First-In First-Out service rules we are taught as children.

What’s especially interesting about traffic, however, is that it isn’t a traditional queue.  The number of people ahead of us is only part of the equation determining what our what time will be since the ‘service rate’ is dependent on how we act towards each other.  By actively cooperating, we can decrease service time and finish service sooner.

In this instance, I think the Nash Equilibrium as it currently occurs enforces that traffic jams will continue to occur.  As much as we may want to cooperate and reach a better equilibrium, there will always be the feeling that if we don’t cooperate we can make someone else later deal with that one extra car and come out ahead ourselves.

From a public policy perspective, two ideas have been proposed.  One is to send rolling roadblocks of police through those areas at a slightly reduced speed which requires a fair bit of manpower and only seems worthwhile on particularly common or intractable jams.  The second is to adjust down the speed limit before the jam, decreasing inflow and again smoothing out the wave.  The latter while more practical, also assumes that people will obey the speed limit and we all know how often that happens.

The best case scenario may instead be to find some way to better incentivize cooperation.  They have certainly proven effective before in solving congestion problems.

Advertisements

4 thoughts on “Ending Traffic Jams

  1. I’m so glad I could serve as an inspiration for this post.

    Also, it’s worth noting that one of the other possible solutions to this problem is self-driving cars. I took a transportation engineering course in undergrad for one of my “cross-engineering” electives, and autonomous vehicles are currently seen as a huge improvement over the status quo in part because of what you described. Since it’s assumed they will be able to react faster than humans and coordinate between each other, there will be a shorter gap between cars without any of the downsides mentioned in your post. As a result, there will be increased total throughput without having to resort to the alternative policy proposals.

    Considering “Google has said previously that it intends to ready the technology for a marketable self-driving car by 2020” (http://www.reuters.com/article/us-google-selfdrivingcar-team-idUSKCN0W51CH), it doesn’t seem like this solution is too far off.

    Either way, I think the topic is really interesting. I didn’t realize it before taking the transportation engineering course, but there’s actually quite a lot of overlap between the fields of industrial engineering and transportation engineering.

    Like

    1. While I think in the long run driverless cars could help, in the short run I suspect they are likely to be as problematic as normal cars since they will be mixed in with humans and have to react and predict how humans will drive; Which is not always easy since we are notoriously bad at following traffic rules. Moreover, humans tend to be pretty terrible at dealing with driverless cars.

      I’ve also seem some speculation that driverless cars could lead to an increase in cars on the road at least in more urban settings. Don’t want to find or pay for a parking spot? Simple enough, tell the car to just keep driving around the block. The biggest argument I have seen for how they could remove congestion in this setting assumes that we will give up individual car ownership and move to car-sharing model. Personally, I doubt that this will occur among the majority of car-owning suburbanites or if it does, it will be a long time coming.

      Finally, while driverless cars could increase throughput for just the reasons you mentioned, there are still systemic issues that cause ‘traffic waves’ to arise that would continue to occur even with driverless cars, but perhaps requiring higher densities of traffic.

      Don’t get me wrong, I LOVE the idea of driverless cars. I think when they become widespread, they will be one of the greatest inventions, as far as traffic safety, that humans ever devised. I just don’t think that under our current system of ownership and antipathy for public transit that they will greatly improve congestion.

      Like

  2. I have a couple of thoughts:

    (1) A book called Traffic by Tom Vanderbilt discusses some of these issues (both from the logistical and the psychological point of view) and is excellent.

    (2) One of the Freakonomics podcasts discussed traffic congestion. In terms of queueing, reducing the traffic by a little bit can make a huge difference. “The average car occupancy in Los Angeles is 1.1. And if it was 1.3, traffic in L.A. would be eliminated.” Of course, if there was no traffic, we’d all drive a bit more and create the congestion again, so it’s hard to win.
    http://freakonomics.com/2014/09/04/regulate-this-full-transcript/

    Like

  3. I appreciated all of the links you included in your post. They were great supplements to the content.

    I’ve been wondering a lot about the “waves” of traffic since I was in elementary school. As is standard practice in most elementary schools, we used to form lines whenever we traversed the school hallways. When I was in the back of the line, I never understood why I had to wait so long after the first person in line started moving to begin moving myself. This seems closely related to traffic waves. I’m glad there is some science to back up my frustrations.

    Amanda and I were discussing “zippering” recently. In some congested east coast cities, the law of the land is that every vehicle on the highway allows one vehicle from the on-ramp to enter the highway. The result is a line of alternating vehicles forming from the on-ramp and highway, hence the clever, clever name. Zippering allows traffic to continue at a more constant albeit slower rate, thereby reducing traffic congestion. Not adhering to the zippering methodology is grounds for more honks than you can shake a stick at. Zippering may not be as effective as the method used in Stockholm (mentioned toward the end of your post), but it certainly is a creative way for people to cooperate to slightly mitigate the negative effects of heavy traffic.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s