What would you do? (Fun with scheduling)

[Perry]

So we are beginning to schedule routes to our new transit terminal that opens in November.  I drew the straw for Route 402 which runs from Boulder City to downtown Las Vegas.  It's a long, long route that currently has an hourly frequency.  I eliminated the portion between downtown and Meadows Mall in March due to the fact that there was already a route that did that and it wasn't very productive. It gave us the chance to interline at the downtown transportation center to use up excess recoveries.

However, in November a new contractor is going to just operate this route...meaning no interlining possibilities.  As I was looking through the schedule and started playing around with it, I realized that at an hourly headway and a stand alone route, it would be extremely inefficient with every trip that terminates in downtown all morning having over 30 minute recoveries.  I hate long recoveries, I like to keep the bus moving.  It isn't my responsibility to figure out when the operator gets a break...that's the contractor's problem.

So, I started adjusting the times to have it run at a higher frequency and got it to run between every 50-55 minutes and the recoveries are balancing out nicely...16 minutes in downtown and about 10 in Boulder City. Trying to keep the B.C. portion tight because the layover is in the middle of a loop and we don't want people to have to sit on the bus forever. 

This is also why I dislike policy headways.  But it also causes a case of not being able to memorize the schedule.  This is a debate I had with a passenger via email and I basically told him that to make a schedule that everyone can memorize for the sake of not having to read a schedule is inefficient and something I would never want to do.

So, for those that schedule for a paycheck and those that ride buses (I'm curious how the sides differ), would you run a bus that can keep moving and run a little more frequently or schedule it to run hourly but end up sitting at one end for an enormous amount of time on several trips?  It's a question we're asked a lot. "Why don't you run a bus every 30 minutes instead of every 23".  So, you'd rather have it do that than more service?  Plus, if your timebands change due to having to add time due to traffic, no gurantee you can maintain that exact spacing.

[WayneNYC]

Nice topic Perry.  For those who don't know, I'm not a transit professional.  That said, I'm grateful for having friends in the industry as I've learned a lot from them.

First, I'd say scheduling a bus on easy-to-remember times/intervals wouldn't be my first priority.  If it happens to turn out that way, then great for everyone.  I'm thinking that if I had a bus on 57 min intervals, I might consider making it an even 60, but I don't think I'd add (or subtract) significant amounts of time to achieve such a schedule/interval.

One plus I can think of in a long recovery is that if the route normally gets delayed, then on occasion some recovery time could help the bus stay on (or close) to schedule for the next trip.  I know back home in NYC, this was often an issue for operators.  Buses would often get delayed and all of the recovery time (and then some) would be lost.

So, I guess I'm saying that if a route has a tendancy to get delayed, then I'd consider building in longer recovery times.  By the same token, if a paricular routes tends to run on-schedule most of the time, then I'd lean toward shorter (but humane) recovery times.  I guess I'd have to play with the scheduling for the various trips depending on the time of day, take educated guesses, hope for the best and learn from any problems for future tinkering/improvements.

[rideonrulez]

Perry, it depends on the situation. I personally like to make it on a schedule that is easy to remember for the customer. But in this situation, it might have to be the exception. So I have to ask, what is the round trip running time on this route and how many buses are operating it?

[aznboy4305]

In my opinion, it depends on the situation as well.

For high frequency routes that run every 5-15 minutes, I wouldn't "pretty up" the schedule to maintain consistency. That is the level of service where most riders wouldn't even look at a schedule to begin with. Those same riders also wouldn't know the difference if the bus came every 7 minutes, or every 10 minutes, or every 7 minutes alternating with every 10 minutes. In this scenario, the utility of optimizing run/recovery time is far greater too, since any run time or recovery savings you gain by not having to worry about a pretty schedule will be multiplied by 6, 7, 8 trips per hour. Operations/Scheduling will reap a lot of benefits and savings, and it will not make any noticable difference to the public.

However with lighter duty routes that run every 30 minutes to every 60 minutes, I prefer schedules that are polished up and run at even/consistent headways. These are the services where transmitting public information is key, since these riders aren't likely to step out to the bus stop whenever they feel like and wait for the next bus. These riders are more likely to study the schedule before starting their trip. The utility of being able to say "Oh the bus is going to come by at :43 and :13" far outweighs adding an extra 1 or 2 trips. Even if it wastes an additional 10-15 minutes in layover, so be it. At this service level, it isn't worth optimizing recovery time for all 1 or 2 trips an hour at the expense of giving your riders and public relations department a massive headache.

[WMATAGMOAGH]

I'd say the line in my mind for where I'd start to want having a memorable schedule is somewhere between 15 and 20 minutes, and certainly after 30 minutes.  There comes a point where ease in memorizing and knowing the schedule is useful because the route comes so infrequently that passengers aren't just going to go out and wait for the bus to show up.  Also, I think that having the memorable schedule is a higher priority for outside of the rush hour periods.

That said, I understand the arguments for running the route every 50 minutes instead of every 60 minutes and probably, the decision to run every 50 minutes or every 60 minutes is one that would need to be made on a case by case basis.  I think if you are going to run the bus every 55 or 57 minutes, you might as well make it a hourly headway.  But I don't think there are any hard and fast rules for how one does this, and a lot of independent variables come in to play each time.

[btconet]

There is probably no right or wrong answer, but I tend to favor a 60 over a 50, but a 40 or 45 over a 60 IF the same is possible.

As an ex-Operator and long time (and still) transit rider, I don't look at routes in cocoons, so having most routes with headways that can offer more consistent transfer connections across the entire span is a big plus to me. I'd hate to be the rider who JUST misses the bus running on a 60 minute headway, waits the headway, and then JUST misses the connecting bus running on a 50 minute headway.  Those are the "never again" episodes that send people running to buy a car.

I've found that a lot of people who don't connect prefer the memory schedules too.  I used to drive our Saturday 14 when it ran on a 50 minute headway and people would always ask me "When do you come back the other way?" If I answered "4:20pm" for example, I usually got a response like "Oh I won't make that. So the buses come through here every hour at 20 past?"  When I told them they'd have to add or subtract 50 minute intervals from that time to get the times of the other buses, I usually got "Never mind, I think I know someone I can call for a ride."  Seriously.

Memory schedules are certainly not without their ills from the schedule maker perspective, particulalrly when you're running 60 minute headways on a route with a 90 minute or 140 minute cycle, or perhaps worse, a 28, 58, or 87 minute one way running time (meaning NO layover).  I've had to employ a number of interlines and forced pulls just to pull off a number of our memory schedules.

And as Ray said, the tighter your headway, the less need I feel to run pristine headways.  The assumption in these cases is that the demand is so strong, widening from say a 12 minute to a 15 minute headway can have significant ill effects on loading conditions.  Besides, riders on lines of less than a 15 minute headway tend to just go when ready instead of referring to schedules, unless there's a short or branch.

To give a little bit of peer comparison, we've been largely converting to clockface headways in peak, base, and weekend day schedules, but there are some exceptions. 

In Fall '09, I widened the 20 line from 15 to 22 minutes so as to overlay a mid-day 6 atop it at a 22 minute headway to create an 11 minute headway on the shared portion.  This was due to overloading on the shared segment while the ends were much lighter.  22 was chosen as it the only headway that would allow me to use the same 14 buses previously assigned to the 20 between the two lines.  We also had to widen the 1 midday from a 30 back to a 35 because the Operators just couldn't make the 2 hour round trip cycle from Sinai to Fort McHenry, and we couldn't afford another bus to bump the cycle up to 2:30.

In your situation, I'd take a peak at the routes that connect and may feed this line to see how they might interact on 50 or 60.  One thing for sure, this example is a good argument against contracting to multiple operators.

[Perry]

Thanks for all of your replies.  Well, what was strange with the schedule is that I could run it hourly for the afternoons and not have really long recoveries, but in the AM from the beginning of the schedule until about 2PM I tried it at 50 and 55 minute frequencies and I got pretty good results.  The round trip running time really varies.  In the AM, you can do a round trip minus recovery in about 2 hours 18 minutes, so it creates a weird cycle time.  In the afternoons it takes it closer to 2 hours 45 minutes just because the dynamics of the route change and the way rush hours work here, the PM actually has a double rush hour since people are going to work and coming back at the same time. Chris, we use 3 buses on it.

If I kept it at 60 minute headways all day, I was looking at consistent 30 minute recoveries at the downtown terminal and more than 10 minutes at the southern end.  So when I knocked it back to 50-55, I ended up with 14-16 at the north end and 7-10 at the south end.  It has really good On time performance so I'm not too concerned.  I also had three factors I looked at, the timing of it getting to the Nevada State College, which is a busy but not hugely popular stop, transfer connections to another hourly route in Henderson, plus the bus Nellis route which has an hourly branch near a major mall in Henderson, and the timing of trips out of downtown in the evening. 

Right now the biggest complaint is that the bus leaves on the hour so if you get off at 5, you miss it.  It's interlined with other routes so it's hard to move the times much and with college sort of an anchor, we didn't want to mess with the timing. 

I was actually able to keep times meshing with college schedules, meet up with the Route 217 in Henderson at times and make really good connections (even better than they are now) with the Nellis route.  I was also able to push the departure time from the downtown terminal to 5 past the hour so when it gets to the heart of downtown, it will be about 10 after, giving those employees more time to get the bus. 

I also was able to squeeze an additional trip out of the schedule by running more frequency even if it meant only a few minutes.  I just couldn't stomach the really inefficient recoveries and this way I got a free trip out of it since we pay by the total revenue hours and that didn't change because of adding the trip in the middle of the day.  The neat thing was it blended in to our later evening service which I didn't want to touch.

Scheduling is such a dilemma.  I had fun with this one and hopefully itworks out.  I know having a schedule that can be memorized is easy, but it has its severe downfalls and with budgets as tight as they are, we have to look at making the service as efficient as possible.

[Perry]

Fortunately this is only short term with contracting out with another company for the 402.  When the Boulder Highway Green Line starts, we're going to eliminate the segment from part of Henderson to the Galleria Mall, also in Henderson and run an express route operated by Veolia I think from the Galleria mall to downtown.  We'll have the Boulder City portion run by that other company and connect in to the Green Line on the southern end.

The reason we're doing that is as I was looking at the 402 vs. the Green Line, which is BRT, the 402 was still going to be faster so to hinder having a publicity problem with an existing route faster than a BRT route that is costing a lot of money we decided to feed as much into the BRT as possible and separate the 402 from the mix and run that more frequently into downtown, establish more park and ride lots for it and maybe get an extension to the north out of it.

Neither of us in the planning dept. is really thrilled with the Green Line and how it was planned for us...Boulder Highway is a really long corridor and no matter how limited stop it is, the freeway portion of a route is still faster.  I want to tie the 402 in with the Green Line near the mall it's stopping at so there is some inter-connectivity there.

Through the frustrations we go through, I still like that we're given the ability to do what we think is the best.   The Green Line unfortunately was sort of planned by other departments, rather than transit...go figure!