Tuesday, March 4, 2014

A Look at What the Port Washington Branch Timetable Could Look Like

This post was published in 2014 and was a discussion on a series of concept timetables for East Side Access developed by The LIRR Today.  The LIRR published a proposed service plan and draft timetables in 2022.  Read about the proposed East Side Access service plan here. 

Today I will unveil the first of 13 different timetables that will show what the LIRR could possibly do once East Side Access is ready.  These timetables represent what I feel would be the best use of the LIRR's infrastructure, and not necessarily the LIRR's, though I have tried to incorporate the general direction they seem to be heading it into the service patterns.

Today we will start with the Port Washington Branch.  While it's likely not the easiest in terms of scheduling and service patterns, it doesn't have to pass through Jamaica, so it does make things easier in that sense.  However, Port Washington Branch trains not having the option to stop at Jamaica does make things more difficult in a sense.  While the LIRR may not emphasize transfers at Jamaica, you will still be able to switch between trains at Jamaica, so your options are not as limited as they might be out of one particular terminal. With all the branches I tried to evenly distribute the trains to Penn Station and Grand Central as best as I could, but I had to pay special attention to this on the Port Washington Banch, but I will discuss this more later.

Included in the linked .pdf file below is sample weekday and weekend Port Washington Branch timetables that include service to Grand Central Terminal.  Peak, off-peak, reverse-peak, and special/event trains are all included.

The timetables only show revenue trains, not equipment or deadhead moves.  I have planned out each deadhead move, equipment rotation, etc., as I was doing this project, and I have most of those moves handwritten down on paper, but I have yet to insert all of them into the actual excel timetables.  Nonetheless, I tried whenever possible to minimize the amount of deadheading, since it is essentially wasted crew and equipment hours, but in some cases, it was unavoidable.

Like I said in the introduction post yesterday, each branch's service will start off with some "Key Assumptions" which need to be taken care of in order for this schedule to be practically feasible.  It is not a total blow if any one of these plans have not been completed, and I have designed the schedules to be modular enough that trains can be subtracted pretty easily if the capacity is not ready on opening day.  I have also included some basic "Service Guidelines" that I tried to stick to.

Key Assumptions
  • The LIRR will complete the Colonial Road Improvement Project, allowing an additional 12-car train (for a total of 24 units) to be stored on the siding.  If the LIRR were not able to complete this project (but I see no excuse for them not to finish it, they've only been going on about it for an eternity) you would have to remove trains that originate at Great Neck from the schedule (and subsequently distribute the stops on to the remaining trains).
  • The Port Washington Yard Reconfiguration will be completed to allow for two additional trains to be stored at the yard.  Due to the nature of the single-tracked bottle neck on the Port Washington, you basically get a one-to-one exchange on yard tracks for revenue trains out of Port Washington.  If you cut out one yard track, you have to cut out one train.  If they can finish neither of the tracks, then two trains will have to be cut out of Port Washington, and that can cause some glaring service gaps.
Service Guidelines
  • Service, at a minimum, will be provided half-hourly to all stations along the branch (except during the reverse-peak and late night periods).
  • During the off-peak periods, trains alternate terminals (between NYP and GCT), allowing for a direct ride to each terminal once an hourly.
  • On weekends, train service is increased during the busiest travel periods (approximately 9:00am-11:30am in the westbound direction and 4:00pm-7:30pm in the eastbound direction).  During these periods, train service operates at 15-minute frequencies to Great Neck with half-hourly express service to Port Washington.
  • During the weekday peak period, service along the branch is split into two main halves, trains that operate east of Great Neck to Port Washington and trains that begin or end their runs at Great Neck, turning on the Great Neck Pocket Track.  To increase service during the peak of the peak, additional trains that begin or end their runs at Bayside or Little Neck have been included.
  • During the reverse-peak period, equipment trains that operate against-the-flow from Great Neck to/from western terminals will accept revenue passengers for all or part of their runs.  This allows for the most robust reverse-peak service where possible, while minimizing the time trains and crews spend traveling empty.
  • Additional scheduled service to support events at Mets-Willets Point/Flushing Meadows has been included.  This does not include unscheduled extras that typically run following the completion of an event.

I designed off-peak service on the Port Washington Branch to cater particularly to intra-city travelers who live past the (7) train's Flushing Terminal.  At this point, current ridership levels do not justify providing 15-minute headways to the stations in Queens during the entire off-peak period, but there is reasonable potential for ridership levels to be able to meet the 15-minute headways.

As of now, I only have 15-minute headways on the Port Washington Branch sustained in the "peak" direction (to New York in the morning and from New York in the afternoon).  I have this service pattern set up in this way to allow the equipment for the Great Neck turns to come from Port Washington Yard on Saturday mornings (saving the expense of emptying the yard on Friday evenings and allowing for the more efficient use of equipment by not having tons of it spend the weekend in Port Washington (annoying all the NIMBYs).  In the same way, eastbound service on Sunday evenings will refill Port Washington Yard for the Monday morning rush.  I have elected to not schedule these trains to continue to Port Washington due to the single-tracked nature of the line.  These last of these trains, after completing their runs, will wait on the Great Neck Pocket track for the opportunity to sneak up the branch to Port Washington, since trains will not be able to proceed east immediately after ending their runs in Port Washington.

During all other off-peak periods, half-hourly service with alternating terminals should be able to provide sufficient service levels during those periods.

In the peak periods, train service is broken up into zones, much like it is today.  Trains that originate in Port Washington tend to only serve the four stations in Nassau County, where ridership is the heaviest, while bypassing all of the stations in Queens which see fewer riders.

At Great Neck, I tried whenever possible to have two trains that depart within a few minutes of each other operate to different terminals.  Take the AM peak on the timetable linked above (page 2).  Train 419 (the 6:37am train from Port Washington to Grand Central) stops at Great Neck at 6:47am.  Immediately following that train is train 221 (the 6:50am train from Great Neck to Penn Station).  This train leaves Great Neck three minutes after the Port Washington train passes through.  So, while these connections will not be guaranteed, it is theoretically possible for someone from Plandome to transfer trains at Great Neck to the local to Penn Station that follows behind the express, essentially decreasing headways from the stations east of Great Neck to either terminal since you do not have to wait for a train that is going to your desired destination.

Like I said above, the fact that transfer opportunities are limited on the Port Washington Branch due to the fact it bypasses Jamaica makes things a little more difficult.  While I tried to do things like coordinate unofficial connections at Great Neck and the addition Woodside stops to more trains give riders more options to get to the terminal of their choice.  It's not perfect, and it is difficult to get the desired headways and frequencies with the very limited capacity on the Port Washington Branch, but I tried to schedule everything to avoid any glaring gaps in service.

Reverse-peak weekday service includes a number of "revenue deadheads" that represent trains that turn at Great Neck.  While scheduling deadheads reduces flexibility slightly, as deadheads that don't accept passengers have a little more freedom in their schedule to get from point A to B.  The PM reverse-peak service includes fewer of these trains than the morning schedules so it can be easier to recover during service disruptions (in the morning, the LIRR has to make the effort to get the equipment to the eastern terminals so service can speedily resume once things improve, but in the evening rush, canceling eastbound trains can be done more easily, though this jeopardizes reverse-peak service).  However, a number of these trains have been kept in the schedule as Mets Extras when the Mets play evening games at home, though they are subject to being annulled during service disruptions.

At this point, I think I've given a basic run-through of what Port Washington Branch service could work like.  Feel free to comment and question anything you might notice.   Time does not permit me to write about, or for you to read, an in-depth analysis of every train, so if something is unclear or you would like to know my reasoning behind sometime, please ask.  Additionally, I have made millions of tiny little changes to this Excel spreadsheet so there might be some tiny (or major) typos that arise, and if you catch any, please let me know of those too.  And suggestions are very welcome.  I tried to think of everything, but I inevitably missed something someplace--I am not fully aware of the needs of Port Washington Branch, so if you see something you'd like to see changed, please let me know.  You can leave a comment at the bottom of any of these posts, (and I have a nifty new comment system to make that even easier to do), or you can e-mail me at patrick@thelirrtoday.com at any time.

View an index of all of the other sample timetables, the posts about those timetables, and revision history, right here.

19 comments:

  1. Amazing work you did here. Can't wait to see what you have in store for the branches I ride daily.
    Just a question. Why didn't you want to offer transfers at Woodside to the alternating terminal? It would be nice to see the option if possible. Even if it requires a bit of waiting, it may be more convenient to wait at Woodside a couple of minutes for a different terminal.

    ReplyDelete
  2. Don't you think you're a little under qualified without a background in operations research? Wild conjecture helps no one.

    ReplyDelete
  3. Perhaps you can define what you mean by "wild conjecture." In the process of creating these timetables I carefully reviewed ridership data (by station, by train), I considered current schedules and service patterns, their strengths and weaknesses, and the performance of current trains (after analyzing day-by-day and train-by-train OTP for every weekday train in 2013), so I didn't just pull these out of my rear end.

    I have carefully thought-out reasoning behind every train. Time constrains prohibit me from putting together lavish, 938-page PDF reports for each branch, as I do have a full-time job of my own. Let's see what you can do better, Mr. Operations Research.

    ReplyDelete
  4. Load balancing played a pretty big role in deciding what trains would stop at what stations, particularly the west of Jamaica stations like Woodside, Forest Hills, and Kew Gardens. With all hope, intra-city travel will rise, and we don't want those passengers to swarm already crowded trains at those stations. I have thought of it, and perhaps some additional Woodside stops should be added to Port Washington Branch trains.

    Thanks for the suggestion.

    ReplyDelete
  5. Yes, whatever remnants are left of the Sunnyside station plan would involve a station that is west of the tunnel portals, thus only serving trains going to/from New York Penn.

    ReplyDelete
  6. 98% of us who like trains just like them for fun. We read this site (and presumably Dr. O'Hara maintains this site) just as a hobby. I could care less for "938-page PDF reports," as I hardly ever get past the table of contents.

    I applaud Patrick's efforts and I think it's fantastic that someone other than the LIRR is taking a fresh look at things.

    ReplyDelete
  7. Try again. OTP is not an optimization strategy by itself. You pulled the numbers out of your ass. For someone who spends so much time dictating what the MTA should be doing, big on words and short on technical competence. Stick to your medical practice Mr Med School Last In His Class.

    ReplyDelete
  8. Where's your "operations research," big shot?

    ReplyDelete
  9. Where's your "operations research" Clifford? I'd love to hear about all of your wonderful research.

    ReplyDelete
  10. A great attempt, Patrick! However, I would look at the hourly gaps in service reverse-peak. On many transit systems, peak-hour, peak-direction ridership is stagnant, while reverse-peak and off-peak ridership is substantially expanding. Additionally, reverse-peak transit use tends to be lower income and without access to a car compared to peak-direction ridership. Three scheduling options for maintaining at least half-hourly reverse-peak service follow. Option 1 can be implemented with no new construction, while options 2 and 3 require a siding on the west side of the tracks south of Plandome Station, from Stonytown Road to Bridge Road.

    1. Maintain half-hourly service in reverse-peak direction, schedule peak-direction trains accordingly. No new line tracks or sidings (yard tracks may be needed). For PM Peak:
    WB: Leave Port Washington at :10 and :40; arrive Great Neck at :20 and :50.
    EB: Leave Great Neck at :22, :25, :28, :52, :55, :58; Arrive Port Washington at :32, :35, :38, :02, :05, :08.
    With no line infrastructure improvements required, peak has six trains per hour, while reverse peak has two trains per hour. Service is a 3:1 ratio of peak to reverse peak trains.

    2. Maintain half-hourly service in reverse peak, construct siding at midpoint south of Plandome.
    WB: Same as above
    EB: Same as above, but with two trains departing Great Neck at :10 and :40, meeting the existing reverse peak trains on the siding.
    A siding is required for this to work, preferably 1/2 mile long so trains do not have to come to a complete stop. Service is a 4:1 ratio of peak to reverse peak trains.

    3. Use a *20* minute headway in both directions, but with additional peak service. The siding in option two would need to be constructed.
    WB: Leave Port Washington at :10, :30, :50; arrive Great Neck at :20, :40, :00.
    EB off-peak: Leave Great Neck at :00, :20, :40; arrive Great Neck at :10, :30, :50
    EB additional peak: Leave Great Neck at :10, :30, :50, arrive Port Washington at :20, :40, :00.
    A regularly-spaced clockface schedule could be operated east of Great Neck, however, only six trains per hour in the peak direction could operate. Service is a 2:1 ratio of peak to reverse peak trains.

    ReplyDelete
  11. And no one forced you to read this post or post abusive comments on it rather than being constructive.

    And where do you think the subway / railway started?

    Yes someone's fantasy that turned into reality.

    ReplyDelete
  12. Thanks very much for your input. I tried as much as possible to sustain reverse-peak service into/out of Port Washington, but when it came time to coordinate the schedules with all of the remaining trains mixing in, it became difficult to add in additional trains (or meneauver the existing ones) without having trains trip over each other going coming out of the tunnels. (This also makes it difficult to implement the "clockface schedule" you proposed.)

    That being said, I did take a closer look at the reverse-peak schedules and I was able to squeak in and additional reverse-peak train in the PM rush hour so the gap in half-hourly service is less severe (essentially, only missing one train). As I make adjustments to the other branches when we get to those I will continue to look at the option of working additional Port Washington trains in when possible, but the limited infrastructure does make it difficult.

    Your comments are still greatly appreciated, and I will keep them in mind as I make other changes.

    ReplyDelete
  13. I like a little fantasy every now and then. But that being said, his timetables seem practical enough. As a former Port Washington Branch rider, I think such a setup could work quite well.


    Until you can come up with a new set of timetables that is as thorough as his, with your almighty "operations research," I suggest you take Chris' advice and take a hike.

    ReplyDelete
  14. I'll second the sentiments of the above commenters. You're more than welcome to suggest constructive criticisms of the above timetables that include your "operations research" (since you seem to know all about it).

    But I will not sit here and let you make ad hominem attacks on me or other persons. You're more than welcome to, and in fact I highly suggest, you do not read future posts if it is going to bother you so much. We'll get along just fine without you.

    ReplyDelete
  15. I think all of the idealists and their "operations research" are the root of many of the railroad's problems. The LIRR isn't some European people mover...it is a complex railroad that has thousands of infrastructure limitations. "Operations research" might be able to tell you the most effective way of doing something, but if you don't know of the infrastructure limitations and how to best work around those, you don't have squat.

    Patrick has shown to have at a minimum a basic understanding of the LIRR's infrastructure, more than you seem to have, so I'll listen to him before some transit idealist who's ridden the LIRR twice, thank you very much.

    ReplyDelete
  16. And how has Patrick's methodology taken account of limitations? Now i can't speak for the MTA, but Amtrak and the freight railroads use OR to design schedules. Their infrastructure is much more dire than the LIRR's. As for why I don't do it myself? I don't work in OR and my experience with it is limited to my undergraduate training in OR, but generally these problems can reach into millions of variables and aren't for the amateurs. A real airline scheduling OR problem used around 500 million decision variables.

    ReplyDelete
  17. You resorted to insults first. As for why I don't do it myself? I don't work in OR and my experience with it is limited to my undergraduate training in OR, and real world problems can reach into millions of variables which are hardly for amateurs. A real airline scheduling OR problem used around 500 million decision variables.

    ReplyDelete
  18. No Clifford you resorted to the insults first. In fact each of your posts is full of them.

    Examples

    "Don't you think you're a little under qualified without a background in operations research?".

    "Is it any wonder that the MTA has no stomach for you "railfans" as you call yourselves"

    As is "Go draw a fantasy subway map, why don't you?"

    And again with your comment " Real world problems ... which are hardly for amateurs."

    As Patrick says if you have constructive criticism then please do give us the benefit of your expertise. Otherwise please stop commenting.

    I read this blog because I (and others) find it interesting. If you don't like it then just stop reading.

    ReplyDelete
  19. @Chris C
    I'm the bad guy for mentioning the standard tool for scheduling which is used by every company in the transportation space. Furthermore given the scale of the changes ESA will have on scheduling, the metrics used by O'Hara are not valid. As for my supposed 4th insult, it's a simple fact that real world operations research problems are usually not simple "toy" examples.

    I'm not bothered if you want to blither at each other, as long as you're aware of the real methodology used. Don't scream at the MTA for not immediately implementing this series of schedules on opening day of ESA.

    ReplyDelete