Emergence of traffic rules

[Robert Holmes-2]

We've been discussing traffic behaviour and the emergence of laws, so here's
a concrete question to consider:


   - When drivers are queuing to get through a lane-restriction (two
   lanes going down to one lane at roadworks, for example) why is the behaviour
   of US drivers radically different from UK drivers?

In the UK, drivers maximally pack the lanes - the merging of the traffic
happens at the last possible moment. In the US, the merging happens way back
from the obstruction. In the case of the jam I was in on I-25 yesterday
(which prompted this post), the traffic was starting to merge about a mile
back from the obstruction and one lane was clear a full half-mile before the
obstruction.

Maybe it's because I'm a Brit, but I think the UK approach is much the
better. Traffic may not negotiate the obstruction any faster, but the
geographic length of the jam is shorter. This impacts anyone who wants to
use an exit before reaching the obstruction.

Also in the UK approach, you don't need to make any (stressful) social
calculations. You stay in lane and when you get to the front of the queue
you take turns merging. In contrast in the US approach, I find myself
driving along the needlessly empty lane thinking "OK, exactly when should I
merge? If I go to far in this lane people will resent me and just not let me
in". In my example on I-25, the decision was made for me because some guy
with a pickup and a horsebox decided to straddle both lanes preventing
anyone from passing him on the inside.

So why did these two different behaviuors emerge? And which truly is the
most effective?

Robert
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20060422/b8e2f3e4/attachment.htm

[James Steiner]

This also depends on the characteristics of the local population.

Optimally, the two lanes should "spread out" leaving 1.5 car lengths
between each car, the gap in one lane opposite a car in the other
lane. Then, the two lanes can merge seemlessly into one. that never
happens. People always bunch up.

Another behavior I see here in the Philadelphia area is that some
drivers will see a "lane closed ahead sign" and immediately merge
over, while others will see that sign, and the open lane ahead of
them, zoom on ahead until the lane finally,ends, and only then try to
merge, having placed themselves ahead of the folks who merged earlier.
 They are viewed with contempt, like any queue-jumper.

This conv. reminds me of this oldy-by-goody, that I'm sure you're all
familiar with:

http://amasci.com/amateur/traffic/traffic1.html

(TRAFFIC WAVES, William Beaty, 1998, Science Hobbyist)

~~James
_____________________________________
turtlezero.com -- its turtles, all the way down!

On 4/22/06, Robert Holmes <rholmes62 at gmail.com> wrote:
> We've been discussing traffic behaviour and the emergence of laws, so here's
> a concrete question to consider:

[Steve Smith]

> When drivers are queuing to get through a lane-restriction (two lanes
> going down to one lane at roadworks, for example) why is the behaviour
> of US drivers radically different from UK drivers?
> In the UK, drivers maximally pack the lanes - the merging of the
> traffic happens at the last possible moment. In the US, the merging
> happens way back from the obstruction. In the case of the jam I was in
> on I-25 yesterday (which prompted this post), the traffic was starting
> to merge about a mile back from the obstruction and one lane was clear
> a full half-mile before the obstruction.
This is not universal in the US.  In the Bay Area, where I recently
spent 1 year, the "UK" approach is more familiar, though perhaps not as
"jammed".
>
> Maybe it's because I'm a Brit, but I think the UK approach is much the
> better. Traffic may not negotiate the obstruction any faster, but the
> geographic length of the jam is shorter. This impacts anyone who wants
> to use an exit before reaching the obstruction.
I agree that this is one consequence.

Another is that some percentage of "defectors" in this prisoners
dillema still exist, racing past the line of hundreds of cars to then
negotiate an entry into the flow, slowing things down "yet more" and
irritating everyone.
>

> Also in the UK approach, you don't need to make any (stressful) social
> calculations. You stay in lane and when you get to the front of the
> queue you take turns merging. In contrast in the US approach, I find
> myself driving along the needlessly empty lane thinking "OK, exactly
> when should I merge? If I go to far in this lane people will resent me
> and just not let me in". In my example on I-25, the decision was made
> for me because some guy with a pickup and a horsebox decided to
> straddle both lanes preventing anyone from passing him on the inside.
This is always an entertaining experience when this happens.  It seems
both Rude and Appropriate at the same time.

>
> So why did these two different behaviuors emerge? And which truly is
> the most effective?
I find that waiting until the last-minute to "merge" is not optimal...
the point of mergeing becomes a false-bottleneck, slowing traffic much
more than is neccesary.  You might notice that as soon as this
bottleneck is passed, traffic speeds up considerably, becoming overly
sparse, the single lane is not being used to it's capacity.

By mergeing early (in-theory), one hits the single lane at it's
"optimal speed" and only minor perterbations occur.    Even though the
merge point (on I 25 in this region in particular) is almost always
*waaaay* too early,  the theory still holds, if the conditions on the
single lane that is the ultimate "real" bottleneck support, say 40 MPH
driving, then the cars in that region will, in fact tend toward that
speed.

The long line, despite the noted negative effect of tying up exiting,
does provide a good strong signal to those who know alternate routes to
take them.  I often take the first exit I see which I know will lead me
to a "surface road" around the jam... like into bernalillo and down
113, etc.   You could consider this another positive optimisation.

The biggest down side, in my opinion, is the "awkward social
decisions".  I too find myself  playing the role of "defector" by
accident, speeding past stopped traffic, trying to figure out when to
merge and then being irritated by those who refuse me entry, when I
chose to "cooperate" before I was forced (front of line, or blocking
merger) to.

When faced with someone wanting into line after "leapfrogging", I am
almost always generous, assuming it was unintentional or at the very
least, not functional to be rude about it.

I suspect we will find in this discussion that there are "two kinds of
people", similar to "cat people" vs "dog people".     I also suspect we
will find that "optimal" is quite slippery, that most of us will not
agree on a single fitness function.   If such functions can be modeled
as linear combinations of agreed upon criteria, we will have different
weights.  I suspect they are not simple linear combinations...

Let the debate begin!

  - Steve