Bring out the riders. Tonight was a mighty toasty 80+ I'd guess. The #9, down to a frequency of every 30 minutes had a full load of about 38 people aboard. My guess, everyone is wondering around out in the heat since we only get about 2-6 weeks of this heat at the most. Personally I'm hoping on only a few more days of this incessant temperature.
This brought up the ongoing debate us transit nerds have been having time and time again, that bus service needs bumped up and paid attention to in Portland. Yeah, we have a world class system, this is true. But I am afraid I must agree, that bus service does truly need some attention. I'm not saying a lot, but a little TLC would go a long way.
Tonight is a prime example, and I know the union isn't able to handle this type of service level, but TriMet should have been able to dynamically bump up service levels within 30 minutes to get an extra bus into the flow out here. Sure 38 people isn't that many, but there where more, as I confirmed with a friend, standing a waiting within another 20 minutes or so. Unfortunately for them that left another 40 minutes until the next bus. TriMet needs to add some dynamic abilities to their services. There are several steps that need to happen.
- The Unions and TriMet need to come to a flexible work shift agreement to bring drivers in on short notice without getting slammed hard for the hours. There really isn't any point in providing the service if the Union requires excessive pay, especially when they could get drivers to volunteer to be available and would probably be happy for the regular pay rates for the time.
- This is the big one, more so than the Union agreement. TriMet needs to be able to respond flexibly to changes in ridership movements. When there are crowds outside that want to ride, TriMet needs to have a way to track and dynamically flex routing muscle when need be. Currently they don't have that capability, at least not by what I could figure just by thinking through some of the issues from a technological perspective.
I'll elaborate on this second point. With a few observable cameras, that are most likely already in place, and a single observer they could monitor flows of people all over the city at key bus stop locations. Based on that they could call in drivers, or drivers currently ready for deployment, to key routes to get people moving. Tonight TriMet has had an opportunity to move several thousand extra people in the course of a few hours, but service levels just remained the same as the schedule dictates.
In today's Internet age, this is kind of silly that service can't dynamically reflect increased flows of people.
I'll write up some more movement and routing ideas in the near future. Some of the other points I've been thinking about is how to improve the bunching that happens on so many routes. I suspect TriMet could gain a 2-5% ridership increase and decrease buses on the road by about 2-5% also just by a few fixes to bunching. But more on that later, for now I'm off to some more code, transit, and general geekdom.
Transit Sleuth 
I am also seeking for a solution for what you say in No. 2. I read about this theory on an urban transit Usenet group, and I think it may be applicable to some agencies.
It’s called "dynamic scheduling".
Bus drivers would hate it, because it represents a break from decades of operating custom. However, they don’t lose any pay, and might actually gain a real lunch break.
Presently, operators bid for their running assignments based on seniority. Typically, their "paddle" or run sheet will have them do laps on a single route. The paddle tells them what time they report to work, what time they pull out, and what times they have to be at their stops. It will also tell them what time they go home.
Under dynamic scheduling, the drivers will only pick what time they start, what route they’d like to begin their day with, and their choice of an hourly or two half-hour lunch breaks. Seniority rules would still apply, and would now be in effect for limiting hourly breaks.
Once they are in service, though, they would be dispatched to the route that needs the next bus to leave — not necessarily going back on the same route they entered. So if a bus enters downtown, a supervisor or dispatcher would look at what buses arrived and put them on the route where a bus would most be needed to maintain headways.
For instance, take Line 12 — a route that serves Barbur on one end of Portland and then in downtown, crosses the river and continues on Sandy. Is this route combination deliberate? Are there really Barbur riders who need a single bus to continue on Sandy? If not, break up Line 12. There will still be a Barbur bus and a Sandy bus, but not on the same line.
The Barbur driver comes in to downtown, but the supervisor says that there is a late bus on 17-Holgate. We need that Barbur bus on Holgate so the late bus doesn’t snowball. The late driver will be dispatched to another line to keep schedule, and so on.
You need a few areas to platoon buses, and it would help if much of the existing service grid could be preserved.
The biggest change in the operating environment is that recovery/layover time as drivers know it would not exist. If a driver gets a 10-minute recovery now, they lose it. If a bus is needed in 4 minutes to maintain a headway, they’ll get sent to that line. It would be a first in-first out system, ideally.
This is where the lunch breaks come in. Since drivers won’t get the rest they are used to with layovers, they will now get either a straight hour or two 30-minute breaks to unwind.
To fill the gap, more of Portland’s reserve fleet would be used. Obviously, when drivers are given breaks, an extra driver must fill in. These extra drivers could also be deployed to fall back if a route becomes critically late.
Excellent follow up comment on that Wad. It would truly be an excellent thing to implement. I’m not sure if drivers would hate it or love it, hard to say. I know generally though, human beings work for change but hate it when they get it. 😉
Either way, whatever they have to do, TriMet really needs to modernize their operating practices. So many parts of the way they operate are not flexible, and are steeped (as you point out) in age old operating customs.
…anyway, I’ll have more follow up on this. It would be nice to actually get service when usage might spike. No reason to miss those opportunities.
I can buy your "point 1" argument.
Some of the work rules about who gets to work are absurd, no doubt.
I don’t really understand what your trying to get at in "point 2".
Wad’s ideas are excellent.
The reason TRIMET has these absurdly long routes is to MAXIMIZE PRODUCTIVITY.
Something Adron would love I’m sure.
Naw… actually I think some of those long routes need split, they are silly. The #12, #9 north to Saratoga, etc. Nobody rides the #9 from Powell, and then up to Saratoga. Even the #4 could stand to have split north and southern sections. Kind of like how the #10 was split into the southern stretch which stayed the #10 and the northern stretch which became the #73.
As for point #2, what I’m saying is TriMet needs to have a way – especially since they have all those stupid useless security camers – to monitor where there are traffic (i.e. lots of people waiting) build ups and dynamically put extra buses onto routes if need be.
Such as if they haven’t noticed an event is happening, and then all of a sudden there is a 1200% increase in bus demand (or light rail) into or out of the city. TriMet should be able to, within 15-30 minutes have another bus or two, or light rail vehicle or two in and running at the point of congestion. TriMet does it for BIG planned events, but they have horrible non-existent response to events that they aren’t paying attention to. The crux is, there are a TON of other events, that cause massive spikes over a period of an hour here or an hour there that would benefit greatly if TriMet could get another vehicle or two into service on short notice. I know they have the ability to do it, they just have to actually pay attention to the city on a daily basis.
Al, I am shocked that my idea would be complimented by a driver.
Seeing your responses, as well as Adron’s, I am thinking those lines are drawn mainly to cut down on layover times and possibly to clear buses out of downtown.
Al, what is the recovery time stipulated in the contract? The Southern California standard is roughly 10% of revenue service, but a minimum of 5 or 6 minutes at each end.
So, dynamic scheduling would also be used to maximize productivity, but in exchange for drivers being sent out almost as soon as they get in, they would get a "real" lunch break.
The other dynamic that would have to change is to have supervisors out at the platooning zones. They should be given laptops and dispatch equipment to ping buses on their locations, as to plan another round of timed transfers.