These buses are bunched intentionally at their Brooklyn Heghts terminal. (Photo by flickr user Stephen Rees)
Last week, when I left my apartment to head to work shortly after 8 a.m., I noticed a common occurrence in New York City: Despite scheduled times that allow for around 6-7 minute headways, the B67 and B69 were traveling north along 7th Ave. one on top of the other. It was a classic case of bus bunching, a problem that plagues many transit agencies and leads to complaints concerning reliability of service. Solving it has never been easy.
This week, an Atlantic Cities post by Eric Jaffe has been making the rounds, and in it, he discusses a new approach to solving bus bunching. Two professors have a “self-equalizing” way of running buses on congested city streets. Jaffe writes:
John Bartholdi III and Donald Eisenstein propose a method of bus coordination that abandons the concept of tightly-managed headways or schedules. Without the restriction of meeting pre-specified targets, drivers instead follow the flow of traffic, and natural headways emerge over time. The result is a “self-equalizing” system with less bunching and more reliability…
The system outlined by Bartholdi and Eisenstein imposes simple adjustments at control points based on an equation whose goal is to both reduce the mean headway on a line and increase overall uniformity of headways. Because the calculation operates independent of a fixed schedule, adjustments can be made at any number of control points. When systems want more headway stability, say for longer routes, all they have to do is add more control points. “The result is that headways are constantly adjusted to become more nearly equal,” the authors write…
For the study, Bartholdi and Eisenstein told drivers to abandon their schedule, ignore headways, and drive with the flow of traffic. When they reached a control point the “self-equalizing” equation computed their next departure time. Compared to average headways recorded on the same day for two weeks prior to the experiment, the new system reduced headway and increased stability
So far, this system has been tested along one 18-mile route in Chicago and on the Georgia Tech campus bus system which sees around 5000 riders per day, and the two say its working. So could this be a piece of the bus problem facing New York City? Ridership is declining because of the unreliability of service. We don’t know when the next bus is coming, and routing needs some work. But solving the bunching problem or at least alleviating it could improve the rider experience, and right now, every little bit helps.
22 comments
I’m willing for the MTA to try anything to make service more reliable. The M15 SBS, which was once a preferable alternative to the subway, has now become a nightmare. Bus bunching is all too common, an every day occurrence. Furthermore, this leads to crowding on one bus but not others, altercations amongst frustrated customers and a bad all-around experience. Furthermore, what once was a 25 minute commute door to door (23rd to Wall) is now on average 45 minutes, following the same route. M15 SBS buses during rush hour average a 10-12 minute turnaround time when they’re not bunched instead of the advertised 6 minute time (South Ferry, not Houston). What was once a great transportation experience is nothing short of awful.
This is off topic, but it caught by attention since I live on the M15 route but have not used the bus in months, though I tried to use the SBS recently and gave up and walked to the subway after a fifteen minute wait.
When the SBS was rolled out, my reaction was that this what the limted service was supposed to be, before the limited service degraded to the level of the regular service. So the cynic in me wouldn’t be surprised to see it degrade to the level of the limited service, though if that happened it seems to have happened awfully quick.
Some of the bus routes could be cancelled or replaced by disabled-only service: eg, buses up Broadway past 59th, where the train stops every 10 feet or so. The bus does basically nothing except provide wheelchair/walker/scooter access.
For most people, the crosstown routes can be replaced by a pair of walking shoes. No bunching on that bus.
“What was once a great transportation experience is nothing short of awful.”
Welcome to the world of transit planners. Most of them are more interesting in polishing some metric they invented and rather enjoy making you miserable. Having or desiring a comfortable trip is elitist, suburban and automobile centric, you know.
Also off topic, but I’ve wondered if the city bus system would benefit from reorganization into an elder division and a commuter division. The elder division would get the slow regular busses and access-a-ride and basically provide transportation to the elderly and disabled. The commuter division would start with the SBS and limited busses.
Given the priorities at the time and the size of the elderly population, service in one or the other division would then be cut or expanded, but unless all the cuts are in the commuter division and the expansion in the elderly division (admittedly likely), the commuter division would eventually provide what is “normal” bus service by the standards of other cities.
So who are all those phantom people who ride those buses you claim do nothing? The crosstown buses run at short headways, many with high capacity articulated buses, and they fill up.
There’s nothing wrong with comfort, but it can be expensive.
Nothing they are proposing addresses the reason for bus bunching.
When a passenger or traffic delays the bus, more riders than average are waiting at the next stop. This increases dwell time, putting the first bus further behind, allowing more passengers to arrive at the next stop, leading to further delays.
Meanwhile, as the following bus gets closer and closer, fewer passengers have arrived at each stop as it approaches, allowing it to get moving even faster until it catches the first bus.
Since SBS allows faster boarding, is bus bunching better where it has been implemented?
Maybe I’m not understanding it, but it doesn’t seem like a solution to me either. It seems to prevent bus bunching but keep the problems associated with bus bunching – or perhaps even exacerbate them further.
Larry’s not describing a solution, but the cause of the problem.
Anyway, the reason off-board fare collection should reduce the problem is that it reduces the extra dwell time per rider, which removes the positive feedback loop that creates bunching.
Reduces, not removes. The subway also falls into the bunching feedback loop even though fares are already collected.
The subway gets into this feedback loop only when crowding levels are so high that extra passengers do make a difference in dwell time.
The subway gets bunched at other times also. It’s just that we don’t notice it. Last Christmas on a Sunday, we wanted to ride the Nostalgia Special but due to fact it was having problems and we thought we just missed it, we kept changing our plans of where to catch it. The result was that we ended up for one hour at 53 rd and 7th watching trains pass. First we waited for the D to go downtown. None arrived for 20 minutes and the E arrived every 5 minutes. When the D finally came, it was already too late for us so we decided to take the E to Queens Blvd instead. Only then, the D started coming back to back, and then there was no E for 20 to 30 minutes.
I don’t want to speak for him, but when he said, “Nothing they are proposing addresses the reason for bus bunching,” he seemed to be saying they basically aren’t doing anything about the problem. The proposal in this blog post seems to rejigger it, but not fix the root problem.
My question is…how much does it work? The data seems to show an improvement in ‘average headway’ of about 30 sec on a 6-min headway route (or, just half that — I may be reading those charts incorrectly). That’s not much.
That said, maybe average times don’t matter — maybe only the worst cases, during heavy/peak load times, matter most, since that is when the most people are waiting the longest? I have no idea how to tell what those numbers are.
That said, if it allows drivers to concentrate more on not running over pedestrians and bikers, then we should implement this plan immediately — were there any actual safety improvements, that would correspond to the drivers’ feeling of having more freedom to worry about safety? It might also allow drivers to protect themselves better from marauding riders, just because drivers can focus on one less thing.
Some other suggestions:
1) Provide safe and pleasant walk and bike environments on bus routes — this way people can hold onto their dignity and get to where they want to go without being dependent on The State to secure their future. There is some truthiness to ‘frequency of motorized transit is freedom,’ but we know that being transit-dependent can only ever provide a small sliver of the freedom provided by human-powered, private transport, where frequency is essentially, perfect/infinity (opportunities to start your departure per second — whereas 1 opportunity to start your departure every 360 seconds is 0.003 Hertz — big difference, which is part of why walking and biking can be so popular, if we allow them to occur).
2) Provide for free public transit, up until congestion pricing is necessary to prevent crush loads.
3) Provide off-bus fare payment.
4) Provide NextBus-like monitors so waiting passengers can know when a following bus is coming.
Keep in mind that Suggestion 1) would allow a transit agency to reduce or cut service on a line much more easily to concentrate on routes that actually should be run, instead of subsidizing development companies and employers, neither of whom are paying their fair share of taxes to support transit. And, no, we should never outsource our public services to private companies, like Transmilenio BRT — privatization is just another name for theft/graft/corruption (yay, BRT and privatization!).
The same thing happens with the BX 29 and BX 23 in the Bronx. Even though they go the same places, they will arrive and leave at main points in their route at the same time. Especially when you’re waiting for 20 or 30 minutes going to and from the Pelham Park 6 Line Station. Don’t know if its bad scheduling or bunching. But it is aggregating.
What about the other bus bunching problem, slow and/or bad bus drivers? Every discussion so far that I’ve read omits this, and apparently assumes an identical level of skill amongst all drivers. This is just not true. So if a route has 8 minutes headways and all buses leave the terminal on time, if bus 1 has a slow driver, and bus 2 does not, then bus 2 will catch up fairly quickly. Compounding the problem is that nowadays, bus 2 will simply follow its leader the rest of the way, meaning that bus 3 will catch up eventually.
The need is to have better bus management, not through some equation, but effective and engaged management. So, OK, dispatchers in the field are supposedly too expensive for MTA, so then with GPS, manage the fleet remotely. Mitigate the problem before it starts, tell bus 2 to pass bus 1, and so on.
In my experience, the slow drivers are usually the ones who are meticulously trying to not run ahead of schedule. Or sometimes they’ve already fallen hopelessly behind schedule (due to heavy traffic or a wheelchair) and, now that they have a handy excuse for being late, are apparently trying to milk it for overtime.
Not only are dispatchers in the field expensive, they also have limited information to work from. GPS is a game-changer.
Let’s just hope the MTA uses GPS to relieve bus bunching. By the way, do you know if there is less bunching on the B63 because of Bus Time? If so, that would be a good indication of its potential?
Also did you see the negative comments about bus bunching on the M15 SBS? Or do you attribute that to the delays caused by the construction of the SAS?
“the B67 and B69 were traveling north along 7th Ave. one on top of the other.”
wow! were they .. like … y’know … doing it?
🙂
Bus bunching is a serious problem but … this just made me laugh! So that’s what they are doing … LOL!
Average headway? That’s just the inverse of frequency. It doesn’t say anything about bunching. If today my 6 bph bus line runs at a nice, even headway of 10 minutes and tomorrow I get three buses every half hour, the average headway is 10 minutes in either case.
If the goal is to reduce bunching, they should be looking at headway variability, not average headway.
It’s not the average headway you should be looking at, because I agree that doesn’t say much. It’s the standard deviation of the scheduled headway at each of the 20 stops along the route that matter and according to the graph under the proposed system there is much improvement.
My problem is that it was tested on a 3 mile loop route. If it is a one-way loop, then it is really a 1 1/2 mile route, hardly comparable with most NYC routes. What would be of greater interest is the 18 mile simulation of a Chicago route which isn’t really discussed.
I’m still unclear exactly how this system works. Is there someone stationed at each checkpoint who releases buses every six minutes regardless if a bus is early or late? If that is the case, how would that work in a City such as NY in terms of paying the drivers. Woudn’t some end up working more than 8 hours while others would finish their tour in less TA eight hours if you abandon schedules as suggested? Somehow I am very skeptical of this system working here.
Bus bunching wouldn’t bother me so much if we had services like in other cities where you can text message the bus to figure out where it is. This way I don’t need to wait in the cold for a while just to see 2 or 3 busses come at the same time. I could comfortably wait somewhere else until its around the block.