Lane closed ahead: Merge now, or at the sign?

[aaronc879]

I used to be a type a road-rager. Then I realized that, if someone felt the need to drive like an idiot, then they probably had to get to something that they at least felt was more important than my safety. Since I'm generally just on my way to work or running errands, I'm in no hurry. Turns out there's just as much work to do at 8:30 versus 8:00.

I'm not advocating excessive speeding or other overly aggressive driving but some people don't have the luxury of coming into work 30 minutes late. 20% of my department was fired last year because of attendance including coming in as little as one minute late. People just need to leave earlier so they don't need to be in such a hurry. Wow, what a concept! It takes me 20 minutes to drive to work with the cruise set at 63-64 in a 65 zone(mileage improved 15% since I reduced my speed 10 mph). I need to be punched in and at a designated place by 4:54pm. I leave my front door at 4:05 a full half hour more than the travel time just in case of bad traffic. I have no sympathy for those who, in my position, would leave at 4:30.

 

[Sarah in SC]

Harry, thanks for chiming in--here's my question: The spot here that "sticks" is a two to one lane coming off a bridge that backs up every day, usually about a mile long line of cars in the "non-merge" left lane, and then handfuls of zippsters coming up on the right lane that ends. This is not construction traffic, or merging onto highway traffic, just a choke point onto an island with (seemingly) no good traffic engineers in the planning department.
What is the "best solution" to something like this situation? Obviously, extending the two traveling lanes would be great, but there are some 100 year old oak trees in the way of that. From a traffic moving standpoint, is the late merge less effective than the early merge? I think I read your paper excerpt to say...it depends.

 

[TromboneAl]

I was having trouble understanding why this was a popular thread, and why it was brought up at all. It wasn't until today that I remembered about traffic congestion. For me, the merge lane means: If there's a slow dude in an old RV ahead, better speed up to 65 now so you can pass while there's a chance.

But I paid my dues with daily commutes across the Bay Bridge. I remember one time in which a woman cut in unfairly. Once we both got on the bridge, I came up beside her, pointed to her tire and mouthed "Flat tire!" She got off at Yerba Buena island. I know, I was mean back then.

 

[RunningBum]

Harry, thanks for chiming in--here's my question: The spot here that "sticks" is a two to one lane coming off a bridge that backs up every day, usually about a mile long line of cars in the "non-merge" left lane, and then handfuls of zippsters coming up on the right lane that ends. This is not construction traffic, or merging onto highway traffic, just a choke point onto an island with (seemingly) no good traffic engineers in the planning department.
What is the "best solution" to something like this situation? Obviously, extending the two traveling lanes would be great, but there are some 100 year old oak trees in the way of that. From a traffic moving standpoint, is the late merge less effective than the early merge? I think I read your paper excerpt to say...it depends.

A late merge is most effective, and it's fair if both lanes are filled evenly and an alternate merge is done. That way there is no place for "zippsters" to unfairly jump ahead of people.

How to train people to use both lanes until the end is another question.

 

[Marquette]

A late merge is most effective, and it's fair if both lanes are filled evenly and an alternate merge is done. That way there is no place for "zippsters" to unfairly jump ahead of people.

How to train people to use both lanes until the end is another question.

But, according to the post Harry just made, a late merge is not that much more effective if it's a 2-1 or 3-2 merge without many heavy vehicles present:

"The results of the simulations indicate that the late merge may not provide as much of a benefit as previous studies had indicated and that the area of application for the late merge may be limited to situations where heavy vehicles comprise more than 20 percent of the traffic stream."

 

[FUEGO]

What is the "best solution" to something like this situation? Obviously, extending the two traveling lanes would be great, but there are some 100 year old oak trees in the way of that. From a traffic moving standpoint, is the late merge less effective than the early merge? I think I read your paper excerpt to say...it depends.

Accept the delay and move on. Or as you have done, select an alternative route that maximizes your utility (ie - one that minimizes your travel time and other travel costs).

From a traffic engineering standpoint, there are some things that could be done to help w/ the bridge problem. One is called "travel demand management" - encouraging smart commuting (carpooling/vanpooling, mass transit ridership, off-peak commuting, flexible schedules, telecommuting, internalizing transportation costs/parking costs, etc).

Another solution is generally called "congestion pricing". I believe the USA has a few examples of this (in California I believe). Basically, charge a toll at the bridge. Make it very low or non-existent during non-rush hour. Then, during the peak periods when congestion is an issue, dynamically set the toll price so that enough people don't want to pay the toll and will take an alternate route or choose to travel at a cheaper time period. For example, if the existing peak hour, peak direction travel demand is, say, 2,600 vehicles per hour, and the maximum capacity of the bridge segment is 2,100 vehicles per hour, then set the toll sufficiently high to dissuade 500 vehicles to select an alternate route. Good luck getting the laws changed to allow this, and then implementing it on an existing facility! It may prove unpopular. SC isn't the most avant garde with their transportation system.

Another potential solution is to "move the bottleneck". Convert from 2 lanes in one direction to 1 lane in one direction at some earlier point. That way, the capacity constraint is moved to a different location and you may eliminate some of the "false capacity" that arises from having a 4 lane road cross a 2 lane bridge.


Good luck w/ that...

 

[rgarling]

But, according to the post Harry just made, a late merge is not that much more effective if it's a 2-1 or 3-2 merge without many heavy vehicles present:

"The results of the simulations indicate that the late merge may not provide as much of a benefit as previous studies had indicated and that the area of application for the late merge may be limited to situations where heavy vehicles comprise more than 20 percent of the traffic stream."

Assuming their simulation is correct, they did show there was no decrease in capacity with the late merge. You still gain the increased queue capacity, increased capacity if there are trucks and control of the zipsters, which would incidentally reduce the blood pressure of the good folks queuing up early.

 

[FUEGO]

But, according to the post Harry just made, a late merge is not that much more effective if it's a 2-1 or 3-2 merge without many heavy vehicles present:

"The results of the simulations indicate that the late merge may not provide as much of a benefit as previous studies had indicated and that the area of application for the late merge may be limited to situations where heavy vehicles comprise more than 20 percent of the traffic stream."

That TRB research paper was really analyzing the throughput at the merge point as the merge scenarios varied. In other words, the throughput at the merge point topped out at about the same capacity for some lane configurations and in other lane configurations the throughput at the merge was increased under the "late merge" scenario.

After a brief skimming of the full research paper it seems that they didn't address directly* the fact that the early merge scenario extends the queue of traffic farther than the late merge scenario. Let's say you have an early merge 1 mile before the late merge point. You have 1 mile of asphalt sitting empty and the queue extends 1 additional mile in the lane remaining open. This isn't a big deal in rural areas with intersection spacing many miles apart. However in suburban and urban freeway systems where intersections are frequently spaced 1 mile apart more or less, that extra mile of queuing might cause serious travel delay and additional queuing at interchanges that would otherwise not be impacted (ie - the queuing might block 9 interchanges instead of 7).

* Note: They did address queue length issues, but not the system-wide effect of longer queue lengths and sub-optimal usage of existing storage capacity. Their analysis was really limited to throughput/capacity at an isolated merge point. Otherwise solid research using state of the practice microscopic simulation modeling (VISSIM).

 

[Danny]

Here in Minnesota we mostly form a line ahead of time - there are usually some though that cut in line - I just chalk it up to the gradual deterioration of spelling standards in school and not getting a good dose of the golden rule growing up.

Although merging is my preference I will squeeze if necessary.

 

[Bestwifeever]

What if there was a law that once you merged into the through lane you couldn't go back into the merge lane? Would you still be an early merger or would you wait until later? (I would wait until the merge lane traffic slowed down.)

 

[Htown Harry]

FUEGO's got the right answers. (Great find on the paper, by the way.)

Sarah, without knowing more about your particular "permanent" bottleneck I really can't say a whole lot about the situation. It may or may not be good engineering.

Although the maximum overall capacity in the corridor is constrained by the bottleneck, the upstream four lane road likely has safety and property access advantages that aren't as obvious as the peak period travel times. Also, there's the downstream network to consider...if the number of lanes on a segment of road doubles, sometimes the result is just a transfer of the peak hour queue to another location down the road a bit. If that's an area has greater density (more driveways, cross-streets, signals, turning movements, etc.) there could easily be an increase total delay for the overall network.

And, sometimes the roadbuilders just gotta do the best they can - where they can - with the dollars, real estate and political support that are available.

 

[Sarah in SC]

Thanks, Harry, you nailed it. The bottleneck would just move further down if they expanded. I appreciate the traffic engineer perspective!

 

[murg]

Here in Michigan the State passed a law several years ago making it against the law to cut ahead of traffic that is merged ahead of the actual merge. I have seen the State Patrol set up teams to ticket speeders and people who cut in late. Nothing feels better than to pass some SOB who blew past everyone else who was following the law and see them pulled over...

 

[FUEGO]

Here in Michigan the State passed a law several years ago making it against the law to cut ahead of traffic that is merged ahead of the actual merge. I have seen the State Patrol set up teams to ticket speeders and people who cut in late. Nothing feels better than to pass some SOB who blew past everyone else who was following the law and see them pulled over...

Seems like a good defense would be "I tried to adjust my speed and merge safely however I was unable to". Unless the charge is similar to a "reckless driving" violation that most states have on the books.

So, in practice, let's say you have signs that say to merge and they are 2 miles before the end of the lane that terminates. But the queue in the remaining through lane is 3 miles long. How do you go to the back of the 3 mile queue when you don't know you have to do so until you are a mile past the back of queue? Do you stop in the middle of the road next to the "merge 2 miles ahead" sign and wait for someone to let you in there? Just curious.

 

[Bestwifeever]

Interesting Michigan merge laws (I'm sure every state has similar trials)

in 2001, sort of as Murg described ("no passing" in merge lane, basically--I don't see how anything other than the no-passing would be enforceable):
MDOT - I-96/M-6 interchange chosen for MDOT pilot project

and in 2005 the evils of the late merge are discussed:
http://www.michigan.gov/documents/MDOT_RR-101_143494_7.pdf

but in 2006, the late merge is enforced (stay in your own lane until the merge point as traffic builds):
MDOT - "Dynamic late lane merge system" will be in effect on I-94 project in Van Buren County starting Monday, July 17

 

[Walt34]

I have seen the State Patrol set up teams to ticket speeders and people who cut in late. Nothing feels better than to pass some SOB who blew past everyone else who was following the law and see them pulled over...

When time permitted we'd do that at one notorious location where the left lane exited (and ended), and the other two kept on going to the often backed-up interstate. The exiting lane was repeatedly and clearly marked (LEFT LANE MUST EXIT) for a mile before but some people would go to the head of the line and cut in.

When one was pulled over, the cheers, thumbs-up signs, grins, and horn-blowing made it clear how the patient ones felt.

 

[HFWR]

In Dallas, people regularly use the service roads as an extra lane, clogging the entrance/exit ramps...