Scientist Says Impatience *Improves* Traffic Flow

http://online.wsj.com/article_print/...974799,00.html

How Brief Drop in Cars
Can Trigger Tie-Ups,
And Other Traffic Tales
July 1, 2005; Page B1

If you plan to hit the roads like the zillions of other drivers this
holiday weekend, Avi Polus has a word of advice: patience.

A transportation engineer at Technion-Israel Institute of Technology
in Haifa, Prof. Polus's concern isn't drivers' collective blood
pressure but traffic flow. Like the growing number of other engineers
and physicists who are hubcap-deep in the science of traffic, he is
determined to explain infuriating mysteries such as phantom traffic
jams (There's no bottleneck or accident at the front of this jam, so
why weren't we moving?) and why a brief drop in volume can,
paradoxically, trigger a long-lasting traffic jam.

Impatience on two-lane roads actually improves traffic flow, as antsy
drivers pass slowpokes rather than letting a convoy form. On highways,
however, "passing, aggressive behavior and lane changing is greatly
detrimental to the flow," says Prof. Polus.
The reason is that chronic lane changing simulates the "weaving
section" of a highway. If an off-ramp lies just beyond an on-ramp,
entering drivers merge left (assuming ramps are on the right) and
exiting drivers merge right, causing traffic to crisscross like mobile
braids. When, in heavy traffic, many drivers change lanes again and
again, trying to find the one that is moving faster, the same weaving
effect kicks in, reducing the capacity of that section of road.

"Weaving is the worst condition for traffic flow," says Prof. Polus.
Because drivers in heavy traffic brake when a car pulls into their
lane, and because it takes time to get back up to speed, there are
larger and constantly-changing gaps between vehicles. That invites yet
more cars to change lanes, propagating a wave of stop-and-go traffic
that cuts the number of cars in a stretch of road by about 10%,
calculates Prof. Polus, who will present his work at the 16th
International Symposium on Transportation and Traffic Theory at the
University of Maryland this month. That may not sound so dire, but in
rush hour the result is a five-mile backup, his calculations show. In
congestion, be content with the lane you're in.

More and more scientists are modeling traffic with equations from the
branch of math called nonlinear dynamics, which describes systems that
suddenly jump from one state to another. Like water that suddenly
freezes, flowing traffic can spontaneously seize up, beginning at a
single point of crystallization (the idiots who braked to rubberneck)
and causing a wave of high density to spread backward.

Lane closures, on ramps, uphill, chronic lane changing and other
"inhomogeneities" in traffic flow can all trigger a density wave,
Martin Treiber of Dresden University of Technology has shown in
mesmerizing simulations (www.traffic-simulation.de/5). One result can
be "phantom" jams, which occur so far upstream of the bottleneck that
the congestion there has long cleared by the time drivers at the back
of the pack reach it. As a result, they never see the snafu that
flipped smooth flow into a stop-and-go mess. By one estimate,
three-quarters of traffic jams are phantoms.

Carlos Daganzo of the University of California, Berkeley, was puzzled
by what highway sensors showed: When congested traffic forms upstream
of a bottleneck, the rate at which cars at the front leave the
congested area decreases. "It's as if, when a line forms at the
popcorn
stand, the server slows down, so people leave with their popcorn at a
slower rate just because there are more people waiting," he says.

Yet the counterintuitive effect is seen time and again, and in a
recent
study he and colleagues figured out why. The congestion causes cars to
jockey across lanes, ever on the lookout for the faster one. Lane
changing increases the gaps between cars, as drivers slow down when
someone barges in front of them. Bigger gaps means fewer cars per
second leaving the front of the jam.

If that seems counterintuitive, consider that briefly reducing volume
can trigger a stop-and-go wave. Within the region with suddenly fewer
cars, perhaps because a long funeral cortege just exited, the emptier
road entices drivers to speed up ("Open road -- yes!"). But sooner or
later, Prof. Treiber notes, these drivers catch up to a denser,
slower-moving region. The ensuing braking can trigger the dreaded
density wave.

Most jams occur way before a road reaches its capacity, and the
culprits are all around you. Even in heavy but moving traffic,
inhomogeneities would have much less effect if drivers had faster
reaction times. When merging traffic causes the driver in front of you
to brake, you do so as well, unless you enjoy fender benders. But
because braking takes time, the gap between you and the car ahead
shrinks, explains Prof. Treiber. You slow even further until the gap
reaches a size you are comfortable with. Result: You are now traveling
even more slowly than the car whose braking triggered the stop-and-go
wave in the first place. The car behind you does the same, and the
effect propagates backward, often for miles.

You can lessen this effect, however. Prof. Treiber suggests looking a
few cars ahead so you know when and how much to brake. "If you brake
just in time, you can usually safely brake less," he says, "which
improves the flow." Consider it a good deed.

"Scott en Aztlán" > wrote in message
...
>
> http://online.wsj.com/article_print/...974799,00.html
>
> How Brief Drop in Cars
> Can Trigger Tie-Ups,
> And Other Traffic Tales
> July 1, 2005; Page B1
>
> If you plan to hit the roads like the zillions of other drivers this
> holiday weekend, Avi Polus has a word of advice: patience.
>
> A transportation engineer at Technion-Israel Institute of Technology
> in Haifa, Prof. Polus's concern isn't drivers' collective blood
> pressure but traffic flow. Like the growing number of other engineers
> and physicists who are hubcap-deep in the science of traffic, he is
> determined to explain infuriating mysteries such as phantom traffic
> jams (There's no bottleneck or accident at the front of this jam, so
> why weren't we moving?) and why a brief drop in volume can,
> paradoxically, trigger a long-lasting traffic jam.
>
> Impatience on two-lane roads actually improves traffic flow, as antsy
> drivers pass slowpokes rather than letting a convoy form. On highways,
> however, "passing, aggressive behavior and lane changing is greatly
> detrimental to the flow," says Prof. Polus.
> The reason is that chronic lane changing simulates the "weaving
> section" of a highway. If an off-ramp lies just beyond an on-ramp,
> entering drivers merge left (assuming ramps are on the right) and
> exiting drivers merge right, causing traffic to crisscross like mobile
> braids. When, in heavy traffic, many drivers change lanes again and
> again, trying to find the one that is moving faster, the same weaving
> effect kicks in, reducing the capacity of that section of road.
>
> "Weaving is the worst condition for traffic flow," says Prof. Polus.
> Because drivers in heavy traffic brake when a car pulls into their
> lane, and because it takes time to get back up to speed, there are
> larger and constantly-changing gaps between vehicles. That invites yet
> more cars to change lanes, propagating a wave of stop-and-go traffic
> that cuts the number of cars in a stretch of road by about 10%,

Whether he realizes it or not, he's just proved that LLBs are the main cause
of traffic congestion, something that most drivers have always known. He
says weaving causes congestion. Without LLBs, there would be no weaving,
just passing. -Dave

In article >,
Scott en Aztlán <newsgroup> wrote:
>
>http://online.wsj.com/article_print/...974799,00.html
>
>Impatience on two-lane roads actually improves traffic flow, as antsy
>drivers pass slowpokes rather than letting a convoy form. On highways,
>however, "passing, aggressive behavior and lane changing is greatly
>detrimental to the flow," says Prof. Polus.

Looks like it depends on the type of road.

--
------------------------------------------------------------------------
Timothy J. Lee
Unsolicited bulk or commercial email is not welcome.
No warranty of any kind is provided with this message.

On Sat, 02 Jul 2005 17:16:57 -0700, Scott en Aztlán
> was understood to have stated the
following:

>
>You can lessen this effect, however. Prof. Treiber suggests looking a
>few cars ahead so you know when and how much to brake. "If you brake
>just in time, you can usually safely brake less," he says, "which
>improves the flow." Consider it a good deed.

"Looking a few cars ahead?"

Is there any other way to drive?

David W. Poole, Jr. wrote:
> On Sat, 02 Jul 2005 17:16:57 -0700, Scott en Aztlán
> > was understood to have stated the
> following:
>
>
>>You can lessen this effect, however. Prof. Treiber suggests looking a
>>few cars ahead so you know when and how much to brake. "If you brake
>>just in time, you can usually safely brake less," he says, "which
>>improves the flow." Consider it a good deed.
>
>
> "Looking a few cars ahead?"
>
> Is there any other way to drive?
>

Heh, yeah, the way most people drive.

That said, it's becoming more difficult these days due to the
proliferation of vehicles whose windows are higher at their bottoms than
the roof of my car...

nate

--
replace "fly" with "com" to reply.
http://home.comcast.net/~njnagel

Around 7/4/2005 12:16 PM, Scott en Aztlán wrote:

> On Mon, 04 Jul 2005 01:41:38 -0400, "David W. Poole, Jr."
> > wrote:
>
>>"Looking a few cars ahead?"
>>
>>Is there any other way to drive?
>
>
> Well, there's "focus on the road 10 feet in front of your car," which
> is how the majority of people drive.

Predicting Judy's response:

"if you deadly criminals would just drive slow like i do in my
beat-to-**** unsafe toastermobile on bald retreads, you wouldn't have to
look more than 10 feet in front of you.? URABLODDTHRISTYKILLER"



--
~/Garth |"I believe that it is better to tell the truth than a lie.
Almgren | I believe it is better to be free than to be a slave.
******* | And I believe it is better to know than to be ignorant."
for secure mail info) --H.L. Mencken (1880-1956)

On Mon, 04 Jul 2005 19:43:00 -0700, Garth Almgren >
was understood to have stated the following:

>
>Predicting Judy's response:
>
>"if you deadly criminals would just drive slow like i do in my
>beat-to-**** unsafe toastermobile on bald retreads, you wouldn't have to
>look more than 10 feet in front of you.? URABLODDTHRISTYKILLER"
>

You coulda forged the headers, and I would have never had known the
difference. Nice job!


--

The last song I started on my PC was: 3 Doors Down - Away From The Sun - Away From The Sun
K:\Audio\3 Doors Down\Away From The Sun\02 - Away From The Sun.mp3
This is track 16 of 457 in the current playlist.

In article >,
Scott en Aztlán <newsgroup> wrote:
>
>"Weaving is the worst condition for traffic flow," says Prof. Polus.
>Because drivers in heavy traffic brake when a car pulls into their
>lane, and because it takes time to get back up to speed, there are
>larger and constantly-changing gaps between vehicles. That invites yet
>more cars to change lanes, propagating a wave of stop-and-go traffic
>that cuts the number of cars in a stretch of road by about 10%,
>calculates Prof. Polus, who will present his work at the 16th
>International Symposium on Transportation and Traffic Theory at the
>University of Maryland this month. That may not sound so dire, but in
>rush hour the result is a five-mile backup, his calculations show. In
>congestion, be content with the lane you're in.

More theoretical posturing from someone who probably never actually
drives in traffic. Maybe he can get together with Leon James of
Hawaii. The situation at a cloverleaf on/off ramp is entirely
different from that on the road in general, because volume typically
changes at a cloverleaf on/off ramp, and there's an additional lane in
play for a short space.

>Carlos Daganzo of the University of California, Berkeley, was puzzled
>by what highway sensors showed: When congested traffic forms upstream
>of a bottleneck, the rate at which cars at the front leave the
>congested area decreases. "It's as if, when a line forms at the
>popcorn
>stand, the server slows down, so people leave with their popcorn at a
>slower rate just because there are more people waiting," he says.

This isn't surprising at all. The cars are slowed by the congestion, and it
takes time for them to get back up to speed, either because the cars
are slugs or people won't put their right foot down..

>If that seems counterintuitive, consider that briefly reducing volume
>can trigger a stop-and-go wave. Within the region with suddenly fewer
>cars, perhaps because a long funeral cortege just exited, the emptier
>road entices drivers to speed up ("Open road -- yes!"). But sooner or
>later, Prof. Treiber notes, these drivers catch up to a denser,
>slower-moving region. The ensuing braking can trigger the dreaded
>density wave.

Except that the lower-density region created by the cars which sped up
creates a buffer zone which absorbs the wave.

>Most jams occur way before a road reaches its capacity, and the
>culprits are all around you. Even in heavy but moving traffic,
>inhomogeneities would have much less effect if drivers had faster
>reaction times.

Inhomogeneities are what prevent traffic from flowing at the
rate of the slowest vehicle. If you study what happens when a road
is near capacity with homogeneous flow and then you introduce a
perturbation, of course inhomogeneity looks like a problem. But
that's not the real world. In the real world there are slowpoke
drivers, large trucks which can't accelerate up hills, people
wandering across lane lines who you don't want to be near, etc. The
additional inhomogeniety introduced by passing these people is a
benefit, not a problem.

On Tue, 05 Jul 2005 10:48:51 -0500,
(Matthew Russotto) wrote:

>This isn't surprising at all. The cars are slowed by the congestion, and it
>takes time for them to get back up to speed, either because the cars
>are slugs or people won't put their right foot down.

That right there is the single biggest waste of road capacity: people
slow down because traffic ahead of them has slowed down, but when that
traffic speeds up again, these numbskull Sloths DON'T SPEED UP! It's
absolutely maddening to be stuck behind one of these incompetent ****s
as the gap in front of their car grows to 5, 10, 20 car lengths while
you're still stuck goping 20 MPH. And people wonder why there is road
rage and freeway shootings...

"Scott en Aztlán" > wrote in message
...
> On Tue, 05 Jul 2005 10:48:51 -0500,
> (Matthew Russotto) wrote:
>
>>This isn't surprising at all. The cars are slowed by the congestion, and
>>it
>>takes time for them to get back up to speed, either because the cars
>>are slugs or people won't put their right foot down.
>
> That right there is the single biggest waste of road capacity: people
> slow down because traffic ahead of them has slowed down, but when that
> traffic speeds up again, these numbskull Sloths DON'T SPEED UP! It's
> absolutely maddening to be stuck behind one of these incompetent ****s
> as the gap in front of their car grows to 5, 10, 20 car lengths while
> you're still stuck goping 20 MPH. And people wonder why there is road
> rage and freeway shootings...
>

The military has understood for decades that leaving a fairly large distance
between vehicles in their convoys (which can stretch out for miles)
eliminates the "accordion effect". They train drivers to leave far more
than 20 car lengths between the other vehicles in their convoy. With that
much cushion, changes in speed along any spot in the convoy (a vehicle
insertion for example) has a very small effect on the following vehicles,
and the effect actually diminishes the further back from that spot one goes.
The opposite situation occurs if the distance between vehicles is too close
(as described in the article).