Throughout Metro’s network, there are a number of common-stop corridors along which multiple routes will run, usually combining for frequent service.  Some of these corridors are scheduled to optimize the distribution of service frequency and eliminate bus bunching, others not so much.  Along the corridors with no schedule coordination, you’re likely to see several buses come at once within the span of a few minutes, then no buses at all for the next several minutes.

A lot of this is the result of other scheduling “hotspots” in the network, or timepoints and pulses, mostly at park-and-rides, transit centers, and other major hubs were infrequent and frequent services come together at one destination.  Scheduling to incorporate both timed connections at these hubs and frequency coordination along common corridors is no simple task– that’s why Metro splits up its scheduling responsibilities by each bus base.

Unfortunately, this brings up another challenge: common corridors are sometimes made up of routes based out of different bases, hence different schedulers.  At Eastgate, for example, routes like the 212 and 218 combine for frequent service to downtown Seattle in the peak*.  However, because both of these routes each come out of two different bases, uncoordinated scheduling often leads to bunching at the freeway stop, limiting the usefulness of passenger capacity the bus at the rear end of a bunch brings.

While an easy administrative solution is to simply apportion the routes along common corridors to one scheduler, I recognize that the network is much more complex, routes need to move around, and common corridors tend to straddle the edges of geographic boundaries.  That said, Metro should develop a standard for developing coordination among separate schedulers for common corridors.

I’m not aware of any other corridors in Metro’s network where this is an issue with all-day frequent service, but assuming the free interchange of routes between bases can happen at any service change in the future, it’ll be a good practice to ensure stable and reliable service along our most heavily-traveled corridors.

*I chose this selfishly to illustrate my personal commute, at the same time recognizing that it’s not a very good example, given the limited peak-only span.  If anyone has a clearer or better example, particularly with all-day service, I’d like to hear it!

41 Replies to “Scheduling and Common Corridors”

  1. Thank you for this very informative post!

    Question: Does Metro do scheduling for the ST routes it operates?

  2. How about the 125 and 120 between downtown and North Delridge? Especially nights and weekends, there are times when each route has 1 hour headways, and the buses come minutes apart. Though with the revised fall service changes, the 125 will be discontinued on weekends, even though the proposed additional 120 service-hours were axed. I guess in that sense, the combined headway problem will go away soon enough.

    1. Once again, the Admiral District has stolen bus service away from higher-ridership Delridge. I hope Council Member McDermott demostrates his social conscience and reverses this terrible outcome.

      1. Jake, I think you missed the work we did on this early last year. It’s actually IMPROVED from how bunched up they used to get. I got Metro to take the schedules from ~85% of 125 runs scheduled at 4 mins. or less apart to only 10%.

        I noticed the improvement, too. The bunching now tends to be more about the 120 falling off schedule in traffic than the runs that are still scheduled less than 7/8 mins. apart.

  3. While waiting for the 255 at CPS in the morning I almost always see a 218 and 212 bunched together.

    1. In the PM 212s are scheduled to precede 218s to avoid crowding. Going the other way would be difficult to precisely schedule due to traffic lights and traffic between Issaquah Highlands and Eastgate but I suspect they try…

  4. It’s a matter of perspective. For the route planners, their job is to move buses, not people. The schedulers have a fixed budget that has to be allocated over a large service area and a long span of service. From a rider’s viewpoint, the most important factor is getting where I want to go. As a rider, I want a bus to be outside the door, waiting at the curb when I’m ready to go. More coordination of schedules is needed and Metro is getting better at addressing schedule coordination problems. For years the 26 and 28 left downtown for Fremont within 2 minutes of each other–one crush loaded and the other nearly empty. But the schedules are very complex and making small changes sometimes have ramifications later in the day that might be negative. Schedulers have a very tough job, but cross base coordination would be very helpful.

    1. If the route planners think their job is to move buses, not people, then they are not in the right job. Metro’s objective is not to move the most buses for the lowest cost, it is to move the most people for the lowest cost.

      And bus bunching increases costs – the heavily loaded bus has longer dwell times at stops and moves more slowly, the empty follower may or may not be able to pass but likely is slowed down. Slower buses cost more.

  5. This is not just a scheduling issue, but also a routing issue.

    For example the 255 and 545 both provide 15 minute service weekday daytimes and 30 minutes service evenings and weekends to common stops at Montlake, Evergreen Pt, and Yarrow Pt. But the 255 is in the tunnel and the 545 is on 4th Ave, so it’s a missed opportunity to have 15 min headway service weekdays & evenings.

    And I believe that I-90 routes to Eastgate operate differently also.

    When routes have common destinations, they should have common originating routing in Seattle.

    1. As incomplete as the tunnel testing has been, I predict there will be more routes moving upstairs. Now is a good time to lobby Metro over which routes those should be.

      1. 255 should be moved upstairs to better serve the SR 520 corridor along with 545, and 554 should be moved downstairs to ease crowding on 550 as Mercer Islanders don’t have to all squuezed into the Bellevue bus and leave the bus half empty only half way thru the route, and also coordinates schedule with the Eastgate/Issaquah commuter routes 212, 217, 218.

      2. Moving the 255 upstairs is realistic. Moving the 554 downstairs is not, due to all the bond issues.

        Realistically, I would be happy to see the peak-only buses move upstairs, leaving the 550 as the only eastbound bus in the tunnel. Long-term, the peak-only buses will have to travel on the surface, and the 550 will be replaced by Link.

      3. Metro/King County built the DSTT. We’re still paying off its bonds. ST bought the tunnel, but Metro is still paying a percentage of the debt based on how many Metro routes vs ST routes are using it. If ST adds routes, it would change the percentage and thus be an additional cost to ST.

      1. I disagree. People shouldn’t have to look through a bunch of schedules before deciding whether to wait for the 255 or the 545.

    2. The 14 and 27 also suffer from similar routing. Two buses running parallel E/W routes less than a quarter mile of flat ground apart. Both buses at 30 min headways, both buses also scheduled within 10 minutes of each other.

      When the 27 was going to be peak-only, I figured the inefficiency of it wasn’t worth thinking about any more. Now I hear the 27 is going to continue to be all-day and the 14 is being reduced to 30 min at peak … madness!

  6. For redundant service on common corridors, I’m going to nominate the Saturday schedule of the 510 and 511 southbound. The 510 and 511 each run every 30 minutes, which means 4 buses per hour. However, the schedule has them all bunched so that if you’re waiting at 45th or 145th, every 1/2 hour, you have a 510 go by, followed by a 511 3-5 minutes later, followed by a 25-30 minute gap waiting for another 510.

    The fact that 510 and 511 have no stops in common north of 145 St. makes things even worse. Last Saturday, I was riding back into town from up north with someone in a car. The driver lives in Mountlake Terrace, so he asked me which bus stop to be dropped off at to get back to Seattle. I had to reply that there were two possible stops, and it would be impossible to tell which one to use without consulting OneBusAway as were were getting close. While I did make it home with no problems – the bus came right on time – the person driving me was left with the impression from this that the bus system is overly complicated and you can’t use it if you’re not willing to spend time pouring over timetables – not the impression we want to convey if we want to convince new riders to use the system.

    (Northbound, the 510 and 511 are, on paper spaced more evenly, however there’s so much random delay getting through downtown that they are still often bunched).

    The same issue used to be a problem on Sundays too, until they replaced the 510 and 511 with the 512. The difference is on Sundays, you get 30-minute headway and we’re only paying for 2 buses per hour. On Saturdays, you still get effectively 30-minute headway even though we’re paying for 4 buses per hour – all to get from downtown to Everett a wonderful 8 minutes faster.

    If a 1/2-hourly 510/511 could be replaced with an every-15-minutes 512, that would be a huge improvement.

    1. “If a 1/2-hourly 510/511 could be replaced with an every-15-minutes 512, that would be a huge improvement.”

      That’s not going to happen because a 512 or 510 requires a larger chunk of service hours than a 511.

      What would be in the realm of feasibility is 20-minute service on the 512 on Saturdays. I’ve been asking ST to do this every time the budget comes up. They tell me they’ll only resort to that if they don’t have enough money to keep the 510 and 511 separate. I don’t get why they place such a high value on keeping the 510 and 511 separate.

      Okay, this is one time that an obvious connectivity improvement isn’t happening because the scheduler in charge is being dense. But I do encourage everyone to keep pushing this suggestion until ST wakes up to the reality that Seattle isn’t the only place riders in Snohomish County are going.

      1. Just make half the 512’s turn back at Lynnwood, then it wouldn’t need more service hours. Or for those who think turnbacks are an abomination, add the 511’s stops to the 510 and schedule them 15 minutes after each other. Then you’d have 15-minute service between Seattle and Lynnwood full time, which is better than 20 minutes.

        The probable reason they want to keep the 510 separate is that Sound Transit was created because people were sick and tired of so many stops, and they wanted a Seattle-Everett bus that didn’t stop in Lynnwood (and a Seattle-Tacoma bus that didn’t stop in Federal Way or SeaTac). But now that we’ve seen that the 511’s stops would add little time to the 510, compared to the major advantage of 15-minute frequency, they should do it now even if it pisses off some Everettites.

      2. Having all 512s and turning half of them back would still require a few extra service hours, due to the extra stops. Reducing the period of 15-minute headway could turn it into a money-saver. But Everett-to-Seattle riders still come out losers due to the slightly longer travel time.

        With the 20-minute headway plan, there is still math to be done on where the break-even point is, but there are no losers. Even the Everett-Seattle riders have the average wait+travel time reduced.

        Either plan is clearly far superior to the existing separated Saturday lines. I don’t know what set of riders the ST planner thinks separate routes are benefitting.

      3. And as a bonus, the 512 could run in the DSTT, since it only runs on weekends, and there are no other Snohomish County routes downtown at that time.

      4. On evenings, keeping the 510 and 511 separate is especially bad because the 510 and 511 each run every hour, rather than every half hour, so keeping them separate means an extra 30 minutes of worst-case wait time and an extra 15 minutes of average-case wait time. And unlike trips to from work, where you can walk out of the office at whatever moment matches the bus schedule, event trips are a lot harder to time because it’s impossible to predict exactly when a movie, concert, or Mariner’s game is going to end, nor is it possible to predict accurately how long it will take to navigate the crowded walkways to reach the bus stop. So frequent service becomes that much more important.

        One thing I will say, though – if you’re going to combine the 510 and 511 with the 510 for more than just Sunday, you have to run the route frequently enough to provide the capacity to carry everyone. I’ve ridden the 512 a few a few times since it started up and even though it runs on an articulated 60-foot bus, it’s packed, with all the seats filled and lots of people standing. So replacing a 30-minute 510/511 on Saturday with a 30-minute 512 would be a non-starter, as it would result in severe overcrowding, at times, with people being left behind. Even with more frequent service, overcrowding could still be an issue if the higher frequency encourages more people to ride the bus.

      5. “Isn’t “the 511 adds miniscule time to the 510″ based on Sundays with less traffic?”

        The additional travel time of the 512 over 511 involves going through 1 transit center, 1 park-and-ride, and one freeway station. All three 511 stops have direct connections to the I-5 HOV lane, so the time cost of these stops is pretty much unaffected by traffic. The only reason it might be more on a non-Sunday is longer dwell times a bus stops due to more people getting on and off the bus.

  7. One corridor of poor interlining waste is Burien to Seatac. The 560 and 180 each provide half-hourly service during most hours, but it is totally random whether they come close to 15-minute headway, or have one trail the other empty. This makes Burien TC all the more annoying as a forced transfer site for everyone from West Seattle and South Park trying to get to the airport without backtracking through downtown.

    West Seattleites still want a one-seat ride to the airport. ST seems to be cowing to their tactics, backed up by wealth. The 560 will never be close to full, while the 120 already is. Having the 120 continue on to the airport would give a lot more West Seattleites and north Burienites a one-seat ride to the airport, with better frequency, and enable the 180 and 560 to terminate at the airport (while also serving Airport Station).

    This would provide much better service for the same cost (albeit less out of ST’s budget and more out of Metro’s). Unfortunately, politics stands in the way.

    1. if the 120 would continue to TIBS or Airport station, I would be pleased. It’s so annoying for such a major bus so narrowly miss a Link transfer.

  8. And now, to venture into the political morass of intra-agency subarea equity preventing decent service for anyone:

    Take a look at the 577 schedule on Sundays: It’s hourly! There is no 578 on Sundays. (When we went through this exercise a year ago, we were expecting 578 Sunday service to be introduced at some point.) The 594 is still half-hourly, as is the 574.

    I’ve gone through the math of how the 574 is actually a slower option for Federal Wayers to get to the airport than the A Route is. So, the only real beneficiaries of 574 service are from Pierce County. I say just make it an express between downtown Tacoma and the airport/station. You’re welcome, Mayor Priest.

    Next, up the frequency on the 594 to 20-minute headway, and throw in the stop at Federal Way TC, so that the 577 can be eliminated on Sundays, and the service hours rolled into extra service on the 594.

    Don’t forget to designate runs that only occur on game days, Seafair weekends, and Christmas rush weekends.
    .

    Lakewooders wouldn’t notice the change from four trips per hour to three, since the replacement would be even headway and lower scheduled maximum time between buses. Yeah, Lakewooders going to the airport would have to transfer in downtown Tacoma. Cry me a river.

    Tacomans would enjoy the increased number of trips, and the increased frequency to Seattle. Federal Wayers would enjoy the substantial increase in frequency.

    I believe this plan, taken as a whole, would reduce service hours and increase ridership. What’s not to like? Well, it might end up charging more to the Pierce County subarea. So, it may be good for every set of riders, but it will be DOA once the penny counters point out that one subarea is gaining financially at the cost of another.

    1. Your proposal sounds great to me. I would even take it a step further. With 594 stopping at Federal Way every 20 minutes, we don’t need an hourly 578 doing that too. So, I would replace the Federal Way stop of the 578 with a stop in Kent, making the 578 a true shadow bus for the Sounder and providing an alternative to the 150’s painful slog through Southcenter. Next, with a faster alternative for people going to Seattle all the way from Kent, the 150 going all the way to downtown Seattle becomes less important. So we can truncate it at Ranier beach and use the savings to boost frequency to the point where we have a timed connection to Tukwila, Kent and Southcenter for every train.

      Going even further, if it were somehow possible to construct a freeway station somewhere around Renton, allowing the 578 to load and unload passengers without getting bogged down in Renton’s traffic and stoplights, maybe a more frequent 578 could even end up replacing the 101, or a least enough of the 101 so what it doesn’t replace can take 106->Link instead.

      It’s amazing what you can come up with, once you lose the idea that as many communities as possible have to have their own separate express bus to downtown Seattle just for them.

  9. A few years ago I remember the schedulers altered the layover times for the 71/72/73, with the result that they all started arriving at the U District within minutes of each other rather than spaced out. I talked to Metro planner and he was pretty pissed off when he found out. From what I understand, one problem is that schedulers are part of Operations and planners are part of, well, Planning, and the departments don’t always communicate well. Schedulers need to be aware of the big picture consequences of their decisions. That said, for some routes it is impossible to space out every segment. You may get perfect spacing in one area but they won’t line up somewhere else.

  10. I remember that there is software available to optimize public transport schedules. Has Metro considered using such?

    1. Yes, that’s what caused the 71/72/73 snafu zefwagner wrote about above. The software could schedule one route but it couldn’t coordinate three routes together.

    1. The downtown transit spine, or Mount Baker Transit Center, depending on which route you leave the neighborhood on.

      1. Spine is different from a hub such as 1st and Mercer, 3rd and Pine, or even University Station. Mount Baker would only work for those close to 23rd Avenue and is further than the the current connections for the #2.

  11. The 43 also has this problem on the corridors it shares with the 8 (John/Thomas) and 48 (23rd/24th/15th). The schedule usually had both coming within a few minutes of each other. It’s particularly bad weekday mornings SB on 23rd, when the 43 and 48 usually travel as a pa ir all the way from where they first Interline at john.

    1. Another reason why I would like to delete the 43, and replace it by doubling frequency on the 8 and 48.

Comments are closed.