Traffic congestion

For the serious discussion of weighty matters and worldly issues. No off-topic posts allowed.

Moderators: Azrael, Moderators General, Prelates

User avatar
BotoBoto
Posts: 191
Joined: Mon Mar 09, 2009 9:31 pm UTC
Contact:

Traffic congestion

Postby BotoBoto » Mon Nov 29, 2010 8:37 pm UTC

Alright, I really do not know where to post this: should we (the world) get education on traffic congestion prevention? I don't see why every accelerates and then brake over and over, and not just try to get a constant speed which prevents traffic congestions on highways.

Does nobody get this? Or is it just me. If this is not good or anything delete please.

General_Norris
Posts: 1399
Joined: Fri Apr 17, 2009 12:10 pm UTC

Re: Traffic congestion

Postby General_Norris » Mon Nov 29, 2010 8:41 pm UTC

How do you know the person in front of you is going to start moving? You have limited vision so you can only start moving when he starts moving. Doing otherwise would mean crashing more often than not.

User avatar
BotoBoto
Posts: 191
Joined: Mon Mar 09, 2009 9:31 pm UTC
Contact:

Re: Traffic congestion

Postby BotoBoto » Mon Nov 29, 2010 8:43 pm UTC

General_Norris wrote:How do you know the person in front of you is going to start moving? You have limited vision so you can only start moving when he starts moving. Doing otherwise would mean crashing more often than not.


No, I mean: keeping a bit more distance and trying to always keep moving, this prevents the "break later, accelerate later" type of congestion a bit.

DSenette
Posts: 2418
Joined: Fri Mar 06, 2009 8:08 pm UTC

Re: Traffic congestion

Postby DSenette » Mon Nov 29, 2010 9:49 pm UTC

as humans, we've been driving vehicles for a hundred years or more. for at least the past 60 we've been doing it en masse. since the inception of the car + road = get somewhere equation we've had traffic jams. the reason? because no matter how smart a person is, if you put them behind the wheel of a car they become morons. add to that all the windows you can look out of when you're driving, and you've got distracted morons.

the only thing that's going to solve the congestion is going to be automated cars that just drive for you and do the traffic management through a hive mind communication system where all the cars know where all the other cars are, where they're going, and when they're going to be there.
The Righteous Hand Of Retribution
"The evaporation of 4 million who believe this crap would leave the world an instantly better place." ~Andre Codresu (re: "the Rapture")

User avatar
CorruptUser
Posts: 10549
Joined: Fri Nov 06, 2009 10:12 pm UTC

Re: Traffic congestion

Postby CorruptUser » Mon Nov 29, 2010 10:27 pm UTC

But then someone will turn off the automation, thinking he/she will go for a joy-ride, and boom, massive traffic pile-up.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7604
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Traffic congestion

Postby Zamfir » Tue Nov 30, 2010 11:12 am UTC

Traffic congestion isn't just caused by people braking too much. There is a real maximum capacity a road can handle at speed X, and if you are around that capacity, deviations (like people braking, or incoming traffic) will amplify into jams.

It's a very complicated phenomenon, but the basic principle is this: the minimum distance between cars on a high way grows at least as fast as their speed, and usually faster. You need some time to react things in front of you, and the distance you travel in that time grows linearly with your speed. In practice, people need and take some extra distance the faster they go. For example because at higher speeds, the differences in speed between vehicles grows too, leading to more complicated traffic.

Think about it: if the distance between cars grows faster than speed, the capacity (in cars per hour) of a road is larger at lower speeds. That doesn't work all the way down to zero of course, but it as a good first estimate for a given road.

So imagine a road close to its maximum capacity at speed X, and some incoming traffic enters from a side road. So locally, the road has to handle more cars than capacity. Which means the cars have to drive slower, and closer to each other. The result resembles a shock wave in supersonic aerodynamics: cars approach at full speed, see the cars in front of them brake, start braking themselves, become part of the slower, denser traffic, and the cars after them do the same. If you look at this from the side of the road, you can see the "shock wave" sometimes as a fixed point in space where every new car hits the brakes.

Now imagine the previous situation, but with so many new cars entering that the road simply doesn't have the capacity at all, at any speed. So after the side road, the speed drops to the speed of maximum capacity, often some dozens of mph below maximum speed. But more cars approach than are put through, and those cars have to drop their speed even lower. They are basically queuing like in a supermarket, and the queue will grow as long as the supply is larger than the road can handle. from the side for the road, you see the "shock wave" of braking cars slowly move backwards, sometimes miles up road.

By the time the original disturbance is gone, you still have this queue to deal with.

User avatar
SlyReaper
inflatable
Posts: 8015
Joined: Mon Dec 31, 2007 11:09 pm UTC
Location: Bristol, Old Blighty

Re: Traffic congestion

Postby SlyReaper » Tue Nov 30, 2010 12:21 pm UTC

Smoother driving as opposed to repeated start-stop is certainly kinder on the mechanicals of the car and the fuel consumption, but it doesn't make a lot of difference to traffic conjestion. Your average speed will be the same.

I've seen experiments done on this. A big circle is drawn on the ground, and multiple people are told to drive at 10mph around this circle. Everybody's speedo is calibrated slightly differently, so although their speedos all show 10mph, some are moving more quickly than others. A faster car closes the gap on the car in front and has to slow down. This forces the car behind him to slow down as well. You can see this bunch-up travelling around the circle in the opposite direction to the direction the cars are travelling.
Image
What would Baron Harkonnen do?

Eowiel
Posts: 166
Joined: Sat Jun 12, 2010 5:57 pm UTC

Re: Traffic congestion

Postby Eowiel » Tue Nov 30, 2010 3:43 pm UTC

I'm not sure to which extent this method also used in other countries, but where I live, when the roads are congested, the police uses a system where they drive some police cars on the highway at a constant speed of 70 km/h and noone is allowed to overtake them, when a sufficient amount of vehicles is following them, another police car enters the highway at the end of the line and starts a new "line". This way multiple blocs of cars are created that drive at a constant speed of 70 km/hours which can speed up traffic a lot.

It's not useful for all sorts of traffic jams though, it's only used when a highway is saturated over a long distance.

User avatar
BotoBoto
Posts: 191
Joined: Mon Mar 09, 2009 9:31 pm UTC
Contact:

Re: Traffic congestion

Postby BotoBoto » Tue Nov 30, 2010 5:27 pm UTC

Eowiel wrote:I'm not sure to which extent this method also used in other countries, but where I live, when the roads are congested, the police uses a system where they drive some police cars on the highway at a constant speed of 70 km/h and noone is allowed to overtake them, when a sufficient amount of vehicles is following them, another police car enters the highway at the end of the line and starts a new "line". This way multiple blocs of cars are created that drive at a constant speed of 70 km/hours which can speed up traffic a lot.

It's not useful for all sorts of traffic jams though, it's only used when a highway is saturated over a long distance.


stopping the traffic entirely is also a good point. then letting everybody smoothyl accelerate

User avatar
Yakk
Poster with most posts but no title.
Posts: 11129
Joined: Sat Jan 27, 2007 7:27 pm UTC
Location: E pur si muove

Re: Traffic congestion

Postby Yakk » Tue Nov 30, 2010 10:04 pm UTC

Short of stopping traffic, how can they start moving at 70 km/h?

The average speed of the pace-making cars cannot pass the average speed of the traffic in front of them. So while this can be used to prevent the traffic from getting worse, it cannot make it any better than the traffic already in front of the police cars.
One of the painful things about our time is that those who feel certainty are stupid, and those with any imagination and understanding are filled with doubt and indecision - BR

Last edited by JHVH on Fri Oct 23, 4004 BCE 6:17 pm, edited 6 times in total.

++$_
Mo' Money
Posts: 2370
Joined: Thu Nov 01, 2007 4:06 am UTC

Re: Traffic congestion

Postby ++$_ » Wed Dec 01, 2010 1:13 am UTC

I think the worst jams are caused by merging, and in the case of merging it really does matter how people behave.

Consider the following situation. A road with two lanes in each direction enters a narrow tunnel. There is only room for one lane in each direction. Therefore, just before the tunnel, there is a merge.

Let's say it's rush hour and the number of cars in each lane is 45% of the capacity of that lane. The left lane is traveling slightly faster than the right lane. Without much loss of generality, assume that the drivers from the right lane have to merge into the left lane in order to enter the tunnel.

Here's what happens: The drivers in the right lane proceed until they reach the point where the lane ends. If there is a space in the left lane, they merge into it. Otherwise, they stop and wait for a space. When a space finally arrives, they move into it. This forces the traffic in the left lane to slow down, because the cars can only accelerate so fast. The slowing causes traffic in the left lane to compress, eliminating spaces that people queuing in the right lane could use. As a result, both the left and right lanes get backed up. There might as well be a stop sign at the entrance to the tunnel.

Here's what would happen if people were eusocial robots: The drivers in the right lane would prepare to merge into the left lane well ahead of time. Because traffic in the left lane is moving slightly faster than traffic in the right lane, they don't have to speed up or slow down in order to do this. As a result, with high probability they would all find spaces well in advance of the merge. The traffic doesn't have to slow down at all.

If the length of time to find places is insufficient (maybe a lot of traffic enters the road soon before the tunnel), then the drivers slow down until they are traveling at a speed that allows everyone to find spaces in time. This is still better than stopping completely.

User avatar
mmmcannibalism
Posts: 2150
Joined: Tue Jun 30, 2009 6:16 am UTC

Re: Traffic congestion

Postby mmmcannibalism » Wed Dec 01, 2010 2:04 am UTC

You example also requires sufficient time on the ramp to merge; even with ideal behavior the road isn't necessarily well designed.
Izawwlgood wrote:I for one would happily live on an island as a fuzzy seal-human.

Oregonaut wrote:Damn fetuses and their terroist plots.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Traffic congestion

Postby morriswalters » Wed Dec 01, 2010 4:40 am UTC

Here is a link to the experiment that SlyReaper is talking about. Here's a British video.

Eowiel
Posts: 166
Joined: Sat Jun 12, 2010 5:57 pm UTC

Re: Traffic congestion

Postby Eowiel » Wed Dec 01, 2010 8:09 am UTC

Yakk wrote:Short of stopping traffic, how can they start moving at 70 km/h?

The average speed of the pace-making cars cannot pass the average speed of the traffic in front of them. So while this can be used to prevent the traffic from getting worse, it cannot make it any better than the traffic already in front of the police cars.


you have to start the system before the road is congested. Since the system is only usable with traffic jams that are caused by saturation and not by accidents that create a bottleneck, it's usually possible to predict how busy a certain road will be. Over here bloc driving is typically used in weekends with nice weather, when a lot of traffic heads to the sea (and back).

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7604
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Traffic congestion

Postby Zamfir » Wed Dec 01, 2010 9:00 am UTC

++$_ wrote:I think the worst jams are caused by merging, and in the case of merging it really does matter how people behave.


It matters, but only in a small band. If the inflow is low enough, it doesn't matter how people merge, and if it is high enough there will be a jam no matter how perfect the cars behave. Smoother merging will push the boundary a bit, but it won't produce magic.

Thing is, if the road can't handle the flow, it will look as if the assholes are the cause of the waves of braking that stop the flow. But without them, something else would have caused the flow to stop.

User avatar
savanik
Posts: 345
Joined: Fri Jan 29, 2010 6:10 am UTC
Contact:

Re: Traffic congestion

Postby savanik » Thu Dec 02, 2010 6:52 pm UTC

++$_ wrote:I think the worst jams are caused by merging, and in the case of merging it really does matter how people behave.


I would like to posit that most traffic jams are caused by poor traffic flow design, rather than people's behavior.

If the length of time to find places is insufficient (maybe a lot of traffic enters the road soon before the tunnel), then the drivers slow down until they are traveling at a speed that allows everyone to find spaces in time. This is still better than stopping completely.


Bad road design can lead to traffic jams quite easily, even with 'good' drivers. Here in KC, we have quite a few highways that run through downtown, originally built when there was no need for today's capacity. Since then, these roads were poorly modified to cope. They didn't have enough space to put in proper entrances and exits without impinging on large buildings or critical city infrastructure. As a result, we have several on-ramps where the space you have to 'merge' is approximately 100'. You barely have enough distance to get up to a merging speed, much less see if the lane you're trying to get into is clear, with enough time to stop safely if it's not. And these merge lanes have no shoulder - if you don't merge, you will pile into a 12' high concrete wall at 55 mph.

Certain locations have reliable traffic issues - the issues are usually related to specific design choices on the part of traffic engineers. Cloverleaf interchanges. Uphill merge lanes without sufficient distance for acceleration by low-powered vehicles. Too many entrances and exits without enough time for proper lane changes between them. And overall insufficient capacity for traffic, especially when combining multiple lanes.
"If it were up to the copyright lobby, owning a pen would be punishable by fines." ---Arancaytar

++$_
Mo' Money
Posts: 2370
Joined: Thu Nov 01, 2007 4:06 am UTC

Re: Traffic congestion

Postby ++$_ » Fri Dec 03, 2010 10:49 pm UTC

savanik wrote:
++$_ wrote:I think the worst jams are caused by merging, and in the case of merging it really does matter how people behave.


I would like to posit that most traffic jams are caused by poor traffic flow design, rather than people's behavior.
I think they are caused by both things.

In particular in the previous post I was thinking about the section of CA-24 near the Caldecott tunnel. There are 4 lanes, which during non-peak hours reduce to 2 lanes. There is always a terrible traffic jam at this point, even though the road is at much less than 50% capacity during non-peak hours. While this is partially caused by the fact that the reduction to 2 lanes is bad design (it's going to be fixed when they add the 4th bore to the tunnel), it is also avoidable, in that if people behaved better the jam would also not be there.

User avatar
El Spark
Posts: 238
Joined: Tue Dec 15, 2009 6:34 pm UTC
Location: Lower nowhere

Re: Traffic congestion

Postby El Spark » Tue Dec 07, 2010 6:12 pm UTC

General_Norris wrote:How do you know the person in front of you is going to start moving? You have limited vision so you can only start moving when he starts moving. Doing otherwise would mean crashing more often than not.


There's this. I'm personally so nervous about other drivers doing something stupid that I treat others on the roads the same way that I would treat your average sociopath with a machine gun: give them plenty of room, and anyone doing anything overtly stupid/crazy/unaware gets the right of way.

Wherever that falls into the various driving models, there you go.
Check out my short horror story "No One Rents 203," available in Kindle and Nook formats.

nitePhyyre
Posts: 1280
Joined: Mon Jul 27, 2009 10:31 am UTC

Re: Traffic congestion

Postby nitePhyyre » Wed Dec 08, 2010 3:55 pm UTC

El Spark wrote:
General_Norris wrote:How do you know the person in front of you is going to start moving? You have limited vision so you can only start moving when he starts moving. Doing otherwise would mean crashing more often than not.


There's this. I'm personally so nervous about other drivers doing something stupid that I treat others on the roads the same way that I would treat your average sociopath with a machine gun: give them plenty of room, and anyone doing anything overtly stupid/crazy/unaware gets the right of way.

Wherever that falls into the various driving models, there you go.

This is called 'defensive driving' and is the way to go.

I'll add a vote that traffic is mainly caused by poor road design. For example, here in Montreal, we have a highway (40 east) that moves along at 100km/h. Except at one portion where the speed limit is reduced to 70km/h. So you have cars pouring in at a rate much faster than is possible to leave. There is never not a traffic jam at this point. Morning rush hour, Sunday midafternoon, doesn't matter. I would love to know how many billions of dollars a year is lost in fuel and lost productivity due to this dumbass decision.
sourmìlk wrote:Monopolies are not when a single company controls the market for a single product.

You don't become great by trying to be great. You become great by wanting to do something, and then doing it so hard you become great in the process.

User avatar
El Spark
Posts: 238
Joined: Tue Dec 15, 2009 6:34 pm UTC
Location: Lower nowhere

Re: Traffic congestion

Postby El Spark » Wed Dec 08, 2010 7:18 pm UTC

People get all offended when I talk about how I drive, for some reason. At least, they do until we compare numbers of tickets and accidents. Just last night, I avoided some moron who was busy running a stop sign by assuming that he was going to run the stop sign. Woot!
Check out my short horror story "No One Rents 203," available in Kindle and Nook formats.

User avatar
Yakk
Poster with most posts but no title.
Posts: 11129
Joined: Sat Jan 27, 2007 7:27 pm UTC
Location: E pur si muove

Re: Traffic congestion

Postby Yakk » Thu Dec 09, 2010 5:04 pm UTC

As you drive through a green light, repeat (out loud, or in your head) as you look left and right:
"looking for crazies, looking for crazies".
One of the painful things about our time is that those who feel certainty are stupid, and those with any imagination and understanding are filled with doubt and indecision - BR

Last edited by JHVH on Fri Oct 23, 4004 BCE 6:17 pm, edited 6 times in total.

jjfortherear
Posts: 103
Joined: Wed Oct 27, 2010 12:44 am UTC

Re: Traffic congestion

Postby jjfortherear » Thu Dec 09, 2010 10:16 pm UTC

++$_ has it nailed. Look up the video "traffic waves" on youtube, by wbeaty.
Qaanol wrote:Actually this could be a great idea. See, you just have to bill the mission to an extrasolar planet as a mission, and then let all the fundamentalists from all religions be the missionaries.

User avatar
c00lguy1337
Posts: 4
Joined: Fri Dec 03, 2010 10:06 pm UTC
Location: Louisville, Kentucky

Re: Traffic congestion

Postby c00lguy1337 » Sat Dec 11, 2010 2:42 am UTC

El Spark wrote:People get all offended when I talk about how I drive, for some reason. At least, they do until we compare numbers of tickets and accidents. Just last night, I avoided some moron who was busy running a stop sign by assuming that he was going to run the stop sign. Woot!

Yep, defensive driving for the win.

I grew up (and learned to drive in) NY, northern NY, in the winter. I've learned that as bad as some NY drivers can be, what I call 'the NASCAR effect' causes a sharp disparity between the number of people who can operate a motor vehicle, and those who can do so socially (IE: drive). Basically, the further south you were born in the continental US, the less capable you genetically are of understanding that automobiles are dangerous tools, and that driving is a boring dangerous chore, and the more likely you are to regard driving as a timed competition, and all other drivers as mere obstacles to overcome. This segues nicely into...

Driving vs. operating machinery. Think that just because you're behind the wheel of a car it means you're a driver? The clubs marked 'W' in your trunk have more right to the name than you do in all likelihood. (Nothing personal, just statistics.) The basics of operating a motor vehicle come easily: gas, brake, steering wheel, buy gas, that's it. The rest of it, the part that involves other people, takes far longer to master, and most people just don't give enough of a shit to bother. Signal lights, All Good Kids Love Milk... oh bother, here it is:
* Aim high in steering - Look as far down the road as possible
* Get the big picture - Maintain the proper following distance so you can comfortably determine the true hazards.
* Keep your eyes moving - Scan, don't stare. Constantly shift your eyes while driving to keep up with changing traffic conditions.
* Leave yourself an out - Surround your vehicle with space in front and at least on one side to escape conflict.
* Make sure they see you - Communicate in traffic with your horn, lights and signals to establish eye contact with motorists and pedestrians.
Now, all of this stuff is EASY, I mean, I can do this stuff, and I'm a low-grade moron. But it does require one thing that your average jackass cannot EVER be bothered to do: plan ahead to leave early enough to arrive on time. So they speed, they do makeup, they eat breakfast, they make phone calls, they do almost everything conceivable BUT drive, while operating their motor vehicles.

Now, added to that, we have the secret monopoly the oil companies apparently have on the companies that manufacture traffic light control systems, allowing them to ensure that programming maximizes wait time at lights for the maximum number of cars (in Louisville, it is customary to see a light go green for an empty side street just as a huge slug of traffic hits the intersection it goes red to). I call this 'the effing oil companies'.

And then we get to the NTSB, and federal highway contracts, and the bailouts. There's literally billions of dollars at stake, and if a state doesn't hold both hands out for its share, the money goes to (gasp, horrors!) some other, less-deserving state.

You put all of this together, and basically, anywhere you put a lot of people, and give them motor vehicles and a road to operate them on, it will go badly. Traffic jams are as unavoidable as syphilis on a two-week bender in Thailand.

morriswalters
Posts: 7073
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Traffic congestion

Postby morriswalters » Sat Dec 11, 2010 5:18 am UTC

Us here good ole boys don't much like how you Carpet Bagger's drive either. :wink: Particularly on I 64.

User avatar
aldonius
Posts: 105
Joined: Sun Jan 17, 2010 2:33 am UTC

Re: Traffic congestion

Postby aldonius » Sat Dec 11, 2010 8:04 am UTC

Build It And They Will Come.

For my part, I would like to posit that traffic increases to fill un-tolled roads into the downtown areas as and when these roads become available.
Corollary: freeways and main roads around any town are almost entirely used by denizens of that town.

juststrange
Posts: 296
Joined: Wed Jul 23, 2008 3:57 pm UTC

Re: Traffic congestion

Postby juststrange » Sun Feb 06, 2011 9:46 pm UTC

Apologies if this is heading toward necro-post territory.

Can we look at this like a mass-flow problem?

Mass flow = Flow Velocity * Density * Pipe Area
Traffic flow = Average speed * (1/Following Distance) * # of lanes

Assuming a static maximum mass flow, average speed is inversely proportional to spacing. The question is, where is the optimum point? Zero cars can travel at infinite speed, and infinite cars can travel at zero speed. Moves same total amount of things.

Heres the kicker. Are you selfish or community minded? The selfish person only cares about average speed, their 'mass' being the only thing that matters. Its how fast you get A-B. The community minded will be willing to work toward highest massflow overall. Again the question: What is the Speed:Distance function, and where is (Speed*(1/Distance)) maximized.

* (I am using 1/FollowingDistance as a stand-in for Cars/Linear Foot)

User avatar
SlyReaper
inflatable
Posts: 8015
Joined: Mon Dec 31, 2007 11:09 pm UTC
Location: Bristol, Old Blighty

Re: Traffic congestion

Postby SlyReaper » Mon Feb 07, 2011 9:36 am UTC

If you're going to use SCIENCE!, you'll need to factor in things like the speed limit and the two second rule, which impose limits on the flow velocity and density respectively.
Image
What would Baron Harkonnen do?

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7604
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Traffic congestion

Postby Zamfir » Mon Feb 07, 2011 10:15 am UTC

juststrange wrote:Apologies if this is heading toward necro-post territory.

Can we look at this like a mass-flow problem?

This has been done to death, and it only works so-so. It predicts behaviour much like supersonic flow, with shocks that stay in one place even while individual cars move through it. There is some truth in that, as first understanding of traffic jams, but it is not accurate enough to really help when designing roads or other measures.

In particular, there are not enough cars on a road to really treat traffic as a fluid. That's especially problematic for multi-lane roads, since mixing between discrete lanes doesn't behave like fluid mixing at all.

And individual road users are much more complicated to model than fluid elements. They do not all behave the same even in the same situations, they behave different at different speeds, at different lanes of the road, near junctions, in different weather and visibility, at morning when they are going to work and in the evening when they are going home, etc. And all those things are relevant

Also, unlike fluid elements, road users have memory and can look forward. So traffic that just decelerated from 70 behaves very different than the exact same traffic situation where the cars are accelerating from 30. And people try to circumvent jams even before they are in them.

User avatar
wbeaty
Posts: 55
Joined: Sun Apr 12, 2009 1:54 am UTC
Location: Bagley Hall basement, futzing with kilowatt VHF pulse amplifiers (UW Seattle campus))
Contact:

Re: Traffic congestion

Postby wbeaty » Fri Mar 25, 2011 1:47 am UTC

Zamfir wrote:And individual road users are much more complicated to model than fluid elements. They do not all behave the same even in the same situations, they behave different at different speeds, at different lanes of the road, near junctions, in different weather and visibility, at morning when they are going to work and in the evening when they are going home, etc. And all those things are relevant


Not true. The myth: driver psychology makes numerical simulation impossible, so it's stupid to even try.

The reality: traffic physicists twerk their Cellular Automata algorithms until they start accurately modeling various "emergent phenomena" which matches the realtime data coming in from road sensors. A research team in Germany did this for the Duisburg-Essen highway network, and then was able to speed up their simulation and reliably predict future traffic jams 30min ahead.
http://www.newscientist.com/article/dn6094-bad-driving-the-secret-to-traffic-forecasts.html

The key here is Emergence and Chaos concepts: often an emergent phenomenon is robust, and will still arise even if the cellular rules at the micro level are altered. It turns out that individual driver psychology doesn't produce enough noise to wipe out the commonly occurring dynamic patterns. But to get it right, the researchers did have to simulate several different driver populations, each following different internal rules. (Older simulations just had cars and trucks. It looks like the newer ones have cars, trucks, ...and aggressive speeders/lane-jumpers.)

I certainly agree that ideal driver behavior would only affect jams but not congestion. Whether a highway section is in "jammed" or "unjammed" state, the number of commuters doesn't change. If a traffic jam is removed, it just means that the totally empty highway downstream of the mergezone bottleneck ...will become full.

On the other hand, people are now noticing something strange going on with US highways:

http://blog.tstc.org/2010/09/23/u-s-traffic-fatalities-plummet-but-why/

http://www.google.com/search?q=%22traffic+deaths+lowest%22

What's the cause? Speculation: during my commutes I've been seeing more and more "gap leaving" drivers among the solid packed columns of cars. Most probably this is the "hyper-miler" revolution, where tens of thousands of drivers are trying to push their cars up above 40MPG by backing off and leaving space. If you maintain a cushion, then you need not hit the brakes nearly as much, yet you still travel at the same average speed as before.

But also you're not killed in a crash involving drivers' reaction times being too short. If we shift the group behavior thresholds a bit, some quantity of fatal accidents turn into near misses.
((((((((((((((( ( (o) ) )))))))))))))))
William Beaty . . .Science Hobbyist page
http://amasci.com . . billb o eskimocom

User avatar
Azrael
CATS. CATS ARE NICE.
Posts: 6491
Joined: Thu Apr 26, 2007 1:16 am UTC
Location: Boston

Re: Traffic congestion

Postby Azrael » Fri Mar 25, 2011 2:16 am UTC

wbeaty wrote: Speculation: during my commutes I've been seeing more and more "gap leaving" drivers among the solid packed columns of cars. Most probably this is the "hyper-miler" revolution, where tens of thousands of drivers are trying to push their cars up above 40MPG by backing off and leaving space. If you maintain a cushion, then you need not hit the brakes nearly as much, yet you still travel at the same average speed as before.

But also you're not killed in a crash involving drivers' reaction times being too short. If we shift the group behavior thresholds a bit, some quantity of fatal accidents turn into near misses.

I seriously doubt that hypermilers are numerous enough to have any discernible effect on traffic. I suspect that it's people (like me) who just don't particularly like getting in traffic accidents by, or because of, crazy tailgating habits.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7604
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Traffic congestion

Postby Zamfir » Fri Mar 25, 2011 5:45 pm UTC

wbeaty wrote:
Zamfir wrote:And individual road users are much more complicated to model than fluid elements. They do not all behave the same even in the same situations, they behave different at different speeds, at different lanes of the road, near junctions, in different weather and visibility, at morning when they are going to work and in the evening when they are going home, etc. And all those things are relevant


Not true. The myth: driver psychology makes numerical simulation impossible, so it's stupid to even try.

All I said was that cars are more complicated to model than individual fluid elements. That doesn't even necessarily mean that traffic modelling is harder than flow modelling, just that a larger part of the modelling effort goes to determining the properties of single elements.

User avatar
philsov
Not a fan of Diane Kruger
Posts: 1350
Joined: Sat Sep 20, 2008 7:58 pm UTC
Location: Texas

Re: Traffic congestion

Postby philsov » Fri Mar 25, 2011 6:00 pm UTC

BotoBoto wrote:keeping a bit more distance and trying to always keep moving, this prevents the "break later, accelerate later" type of congestion a bit.


Not really. At least, not in Houston.

If there is a gap in front of you large enough to fit a car, a car will insert itself there as it attempts to weave through traffic. If you leave a car+ spaced gap in front of you (ie, you slow down or wait or something), then another car will take that spot. Repeat whenever this occurs, and now you and everyone behind you is going 5 mph and everyone in front of you is going 15 mph. The traffic becomes less... equally distributed for the individuals, and the net effect is the same.
The time and seasons go on, but all the rhymes and reasons are wrong
I know I'll discover after its all said and done I should've been a nun.

Eyat
Posts: 129
Joined: Tue Jan 05, 2010 3:48 pm UTC

Re: Traffic congestion

Postby Eyat » Fri Mar 25, 2011 7:22 pm UTC

juststrange wrote:Apologies if this is heading toward necro-post territory.

Can we look at this like a mass-flow problem?

Mass flow = Flow Velocity * Density * Pipe Area
Traffic flow = Average speed * (1/Following Distance) * # of lanes

Assuming a static maximum mass flow, average speed is inversely proportional to spacing. The question is, where is the optimum point? Zero cars can travel at infinite speed, and infinite cars can travel at zero speed. Moves same total amount of things.


Traffic engineering can have some complicated modeling techniques but here is one of the simpler ones.

http://www.cdeep.iitb.ac.in/nptel/Civil%20Engineering/Transportation%20Engg%20I/33-Ltexhtml/nptel_ceTEI_L33.pdf

I learned it in undergrad level traffic engineering courses and it basically describes really obvious things about traffic behavior in a way that appeals to the mathematically minded (wait, are you telling me avg speed goes down when cars per mile goes up?!) There is also the concept of Level of Service which seems to vary from country to country according to wikipedia but as taught to me takes some basic geometrical data like lane width and some traffic data (percent heavy vehicles and the like) and assigns an A-B-C-D-E-F grade to it.

One of the major problems is when designing a freeway basically you expect a D rating during rush hour and you have to guess your traffic data since there is no road there to count. It then takes between 5 and ten years to build the new roadway and in that time a lot of what you based your assumptions on may have changed so you can open a brand new road and already it is obsolete for what the community needs. But you need land to expand it and sitting on that land is a new community.

User avatar
Antimony-120
Posts: 830
Joined: Wed Apr 09, 2008 4:16 am UTC
Location: Wherever you can look - wherever there's a fight, so hungry people can eat.

Re: Traffic congestion

Postby Antimony-120 » Sat Mar 26, 2011 11:43 pm UTC

aldonius wrote:Build It And They Will Come.

For my part, I would like to posit that traffic increases to fill un-tolled roads into the downtown areas as and when these roads become available.
Corollary: freeways and main roads around any town are almost entirely used by denizens of that town.


You've got that in reverse, classic Post-Hoc Ergo Proctor-Hoc. In olden times roads were built because enough people went through the path that it became necessary to build a road. They were also expanded for similar reasons.

In modern times with urban planning we are considerably more able to predict when new roads will be necessary, so we try to have them built in time. If it's done well it looks like the road sprung up just before the people arrived to fill it. Done poorly you get either a road nobody needs, or a road still being built when it was needed yesterday. Boom-and-Bust Towns are good places to see this sort of thing in action (i.e., Look at a number of rural highways built during the 50's that go nowhere, and a number of cities whose infrastructure has been suddenly inundated, like the oil boom-towns of Calgary and Edmonton).

EDIT: Interestingly it looks like what the op is talking about is not actually a traffic jam (though commonly called one), where waves of braking come though. This has an explenation similar to what Zamfir mentioned earlier.

Suppose there's a steady stream of cars on the road, sitting just at carrying capacity. Somewhere in this stream someone has to brake, not because they were a bad driver necessarily, just because somebody cut them off, they were letting someone merge, whatever. It's a relatively minor movement, but it propogates, much like the "queue" he suggested. As a point of fac it might even propogate forwards, although at a slower speed than the cars are going.

So we've got this wave of braking cars. On a road that is slightly under capacity the wave dissipates into the spaces between cars. On a road that is already above capacity, the car behind doesn't have enough time to brake as slightly (his reaction time doesn't leave enough space left to brake slightly) and he has to brake harder, and then the next guy and so on. This then propogates and grows larger, and so on and so forth resulting in a traffic jam and the resulting queue as Zamfir explained.

However, as was mentioned earlier, carrying capacity increases as cars slow down, and people tend to try to avoid dangerous driving. So they slow down until the road is nearly perfectly balanced at carrying capacity. At this point something interesting happens. That first small brake leaves the guy behind just enough time to react and have the exact same small brake. Then the guy behind, and so forth. This creates a wave of brakes (that can move backwards or forwards or be a standing wave, depending on exactly how close the road is to carrying capacity). Now for every little disturbance that causes a brake there's one of these waves floating around. Get enough of them (which statistically speaking you likely will) and you get a series of waves, where you keep having to slow down and speed up.

The important thing to note about this scenario is that it doesn't require an idiot driver. Everyone in the scenario drove with enough space to react to the car in front of them, and nobody broke for a stupid reason. The traffic never stops moving, because the road isn't over capacity (remember what I mentioned about people tending to drive at the speed that gives carrying capacity), but it doesn't move smoothly because there's no spaces left to dissapate these waves.

Naturally in real life where there ARE idiot drivers, the waves show up more often and have an annoying tendency to grow in size until it's less "a series of slight brakes" and more "Dead Stop for 5 seconds, move forward for 10 seconds, stop for 5, forwards for 10, stop for 5, forwards for I WILL MURDER EVERYONE ON THIS GODDAMN ROAD". Which doesn't help, because angry drivers are bad drivers.
Wolydarg wrote:That was like a roller coaster of mathematical reasoning. Problems! Solutions! More problems!


****************Signature Dehosted, New Signature Under Construction************************

pizzazz
Posts: 487
Joined: Fri Mar 12, 2010 4:44 pm UTC

Re: Traffic congestion

Postby pizzazz » Sun Mar 27, 2011 4:20 am UTC

Maybe I missed it, but it doesn't seem like anyone has pointed out that some roads just have too many cars. It was mentioned above that there is a pratical minimum distance that can exist between cars, increasing with speed, and so if there are more cars in the same distance, they will eventually have to slow down. What happens around, NE of NYC, here is that the highways are completely inadequate to handle even normal rush-hour traffic. The Merrit Parkway in particular is 2 lanes in each direction of narrow, windy mess, so of course there are lots of accidents (I don't remember the last time it rained without causing a jam). But even during a nice day, rush hour consists of parking lot speed for 4 hours and there are easily 5 or more times as many people trying to get through as the road can handle. I think much of the Boston-DC corridor has this problem, because a lot of the roads were designed when horse-drawn buggies were popular.

User avatar
aldonius
Posts: 105
Joined: Sun Jan 17, 2010 2:33 am UTC

Re: Traffic congestion

Postby aldonius » Mon Mar 28, 2011 7:16 am UTC

You've got that in reverse, classic Post-Hoc Ergo Proctor-Hoc.

I'm just thinking Induced Demand - traffic growth is typically higher than planned for, and the road reaches capacity before it was planned to.

User avatar
Azrael
CATS. CATS ARE NICE.
Posts: 6491
Joined: Thu Apr 26, 2007 1:16 am UTC
Location: Boston

Re: Traffic congestion

Postby Azrael » Mon Mar 28, 2011 11:38 am UTC

aldonius wrote:Corollary: freeways and main roads around any town are almost entirely used by denizens of that town.

This right here indicates that you've never lived somewhere where commuter traffic is a real problem. Or just real, for that matter. And it's easily debunked (entirely unscientifically) by watching which way traffic moves at rush hour. If it moves into a city during the day, and out of the city at night, it certainly isn't residents causing the traffic.

You'll find that most major cities have traffic patterns like this. In fact, the typical American infrastructure is built as such -- Easier access to cars and the creation of the highway system allowed for this type of flow. It is how and why suburbs exist.

User avatar
cjmcjmcjmcjm
Posts: 1158
Joined: Tue Jan 05, 2010 5:15 am UTC
Location: Anywhere the internet is strong

Re: Traffic congestion

Postby cjmcjmcjmcjm » Mon Mar 28, 2011 2:21 pm UTC

@Azrael: I think ald was referring to metropolitan area. If it is just a town, that wouldn't make sense, but a big enough city can surround suburbs or be big enough that it might be a vacant at night business city and suburb ring, but it is all incorporated into the same city.
frezik wrote:Anti-photons move at the speed of dark

DemonDeluxe wrote:Paying to have laws written that allow you to do what you want, is a lot cheaper than paying off the judge every time you want to get away with something shady.

User avatar
Azrael
CATS. CATS ARE NICE.
Posts: 6491
Joined: Thu Apr 26, 2007 1:16 am UTC
Location: Boston

Re: Traffic congestion

Postby Azrael » Mon Mar 28, 2011 2:49 pm UTC

Sure, if you define a study area to be sufficiently large that all traffic in the study area is therefore from the study area. But that is thoroughly meaningless.

More useful would be the otherwise predefined Metropolitan Statistical Areas, Combined Statistical Areas or even just urbanization lines. Now, perhaps, one can say that the majority of traffic within those zones both starts and finishes within those zones, but certainly not *all*. Also note that all of these groupings are geographically large; to leave Boston's CSA from Boston itself would take 75-125 miles of driving. That's far enough to question the validity of the boundary in terms of commuter congestion. Plus these areas cross town, county and even state lines. Furthermore, they can encompass distinct cities, each with their own suburban ring and with drastically different major traffic flows that have little driver-based crosstalk (i.e. Boston, Worcester, Manchester, Providence).

Using such groupings not only thoroughly trounces The Aldonius Corollary, but from a far more important aspect, complicates funding, planning and responsibility.

(As an example)
Image

User avatar
aldonius
Posts: 105
Joined: Sun Jan 17, 2010 2:33 am UTC

Re: Traffic congestion

Postby aldonius » Tue Mar 29, 2011 7:28 am UTC

Ahh, my "Australian" definitions are perhaps confusing things for me - what I think of as a metro area is the entire built-up stretch from CBD through to the suburbs, and then some.
Could be that I'm overgeneralising my experience - freeways near where I live are dominated at peak hour by commuter traffic.
My "corollary" really is just common sense - "intra-city traffic dwarfs inter-city traffic".


Return to “Serious Business”

Who is online

Users browsing this forum: No registered users and 15 guests