The International Simutrans Forum

 

Author Topic: Passenger capacity in parentheses?  (Read 6435 times)

0 Members and 1 Guest are viewing this topic.

Offline megasycophant

  • *
  • Posts: 18
Passenger capacity in parentheses?
« on: November 12, 2010, 03:42:06 PM »
What does it mean when a number is given in parentheses after the passenger capacity? The AEC Swift bus, for example, lists 42 (20) Passengers as its capacity?

Thanks much,
Brian

Offline inkelyad

  • Devotees (Inactive)
  • *
  • Posts: 482
  • Languages: EN, RU
Re: Passenger capacity in parentheses?
« Reply #1 on: November 12, 2010, 04:26:01 PM »
It is overcrowded  capacity. AEC Swift bus can carry 62 passengers. But it will drop comfort factor. So no extra money for you.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18422
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Passenger capacity in parentheses?
« Reply #2 on: November 27, 2010, 01:44:12 PM »
Reducing the comfort will have more of an effect on longer journeys than shorter journeys, so it won't matter much if your urban 'bus routes are crowded, but you should make sure that your medium and long-distance services are not overcrowded if possible.

Offline sdog

  • Devotee
  • *
  • Posts: 2039
Re: Passenger capacity in parentheses?
« Reply #3 on: November 27, 2010, 06:17:09 PM »
Quote
but you should make sure that your medium and long-distance services are not overcrowded if possible.

this is very difficult, if not impossible to do. If high capacity units, like trains arrive, a high number of passengers arrive at one moment. Increasing the number of units on a line, will cause one bus being overcrowded and the next one empty.

remedies could be:
a 'soft arrival' of pax at a station. They are not immediately ready to leave but have a gauss distributed delay.

a no 'overcrowding allowed' flag to be set on a line.

Offline inkelyad

  • Devotees (Inactive)
  • *
  • Posts: 482
  • Languages: EN, RU
Re: Passenger capacity in parentheses?
« Reply #4 on: November 27, 2010, 06:22:04 PM »
If high capacity units, like trains arrive, a high number of passengers arrive at one moment. Increasing the number of units on a line, will cause one bus being overcrowded and the next one empty.
Exactly like in RL.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18422
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Passenger capacity in parentheses?
« Reply #5 on: November 27, 2010, 06:25:12 PM »
SDog,

neither of the two solutions proposed have any parallel in real life, and, as Inkelyad points out, the problem that they are designed to solve is itself realistic.

Offline sdog

  • Devotee
  • *
  • Posts: 2039
Re: Passenger capacity in parentheses?
« Reply #6 on: November 27, 2010, 10:48:06 PM »
James, i think both have a parallel in real life. The first method is the distribution of time the passengers need to get from one vehicle to the next. Having long ways flattens the peak. (This should be very similar to the broadening of wavepackets in physics). The very nice effect of this is, that a increasing the frequency of vehicles loading will reduce the overcrowding slightly, but has some of the slowest pax left for the next vehicle.

The second suggestion has also application in real live: Reservation only Lines. Examples are Shinkansen and N'EX in japan or Chinese high speed rail. I also have quite some doubts english gentlemen would have accepted to stand in a train at the fin de si├Ęcle.

However, the method i proposed is unnecessarily complicated, and only required for exceptions like the german ICE, where only a few lines are reservation only, most lines can be are very much overcrowded.
A much easier way would be on the pak set size, now overcrowding for long distance luxury vehicles. Perhaps introducing different classes, where the first class can't be overcrowded.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18422
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Passenger capacity in parentheses?
« Reply #7 on: November 27, 2010, 10:55:17 PM »
Introducing different classes would, if done properly, require an entirely new layer of simulation of affordability, which would be an enormous task. One can simply produce similar vehicles designated as different "classes", some with lower capacity (and no overcrowded capacity) and higher comfort, and others with higher capacity and lower comfort (allowing overcrowding), which is a crude but workable solution within the code as it currently stands.

However, I am having some trouble understanding your first paragraph - can you rephrase, perhaps?

Offline sdog

  • Devotee
  • *
  • Posts: 2039
Re: Passenger capacity in parentheses?
« Reply #8 on: November 28, 2010, 02:32:08 AM »
Quote
One can simply produce similar vehicles designated as different "classes", some with lower capacity (and no overcrowded capacity) and higher comfort, and others with higher capacity and lower comfort (allowing overcrowding), which is a crude but workable solution within the code as it currently stands.
This is exactly what i meant. I don't consider it as crude at all, a solution coded in the game itself would not bring any advantages. Even if first class pax would be treated differently from the beginning of their journey, the outcome would be the same for a large passenger network, it just averages out.

Providing only high cost and high luxury service would reduce the number of possible passengers. But this effect should can be neglected in the simulation, we only have a rough representation of passengers anyway. You also have a mechanism in the game that works exactly against this. Car ownership. Higher luxury and cost would reduce the available pax who could not afford it, but draw more affluent pax away from cars. Before introduction of cars there is no alternative way of travel anyways.

Sorry for the first, rather convoluted, paragraph. Here's a second try.

Real live passengers move with different speeds and need finite time to get from one vehicle to the next. Their speed distribution causes the same distribution of the arrival time at the point where they can board the next vehicle. This should be a Gauss-distribution.

So far we considered arriving passengers to be instantaneously ready to board the next vehicle they need to transfer to. It worked well until we had overcrowding. But with overcrowding a non-instantaneous availability of passengers provides a way to reduce peak convoy loads. If service to the station is high enough, some of the early passengers will be fast enough to catch the previous convoy, the majority will overcrowd one convoy and the slowest the next convoy. For infrequent service, all passenger will still try to get into one convoy.

Offline inkelyad

  • Devotees (Inactive)
  • *
  • Posts: 482
  • Languages: EN, RU
Re: Passenger capacity in parentheses?
« Reply #9 on: November 28, 2010, 05:02:50 AM »
You just need big enough bus stop with several buses waiting for wave from next train. (Again, just like in RL. ). With current code bus will not be overcrowded when there is alternative seats in bus from the same line.


Offline sdog

  • Devotee
  • *
  • Posts: 2039
Re: Passenger capacity in parentheses?
« Reply #10 on: November 28, 2010, 05:41:54 AM »
I see how this could work in simutrans, using choose signals (and extensive tunnels). But where is this done in RL? First buses on wait for load and most unusually several buses waiting at the same time. I've only seen (and heard of) a single stop/bay per line where one bus gets loaded after another.

In local transport the overcrowding is not a real problem, it is only a bit strange. Ingame i can just reduce capacity enough for this. It is a problem on the long distance trains. (my first suggestion wouldn't matter for those however)

Offline inkelyad

  • Devotees (Inactive)
  • *
  • Posts: 482
  • Languages: EN, RU
Re: Passenger capacity in parentheses?
« Reply #11 on: November 28, 2010, 06:46:03 AM »
I've only seen (and heard of) a single stop/bay per line where one bus gets loaded after another.
It is really common.
Kazan rail station
Saratov rail station

Krasnoyarsk rail station
Voronezh rail station
Krasnodar rail station
There is multiple buses here.
They are loading one after another. But next bus is waiting near and passengers see it.
Quote
It is a problem on the long distance trains.
IMHO Long-distance trains must not have overcrowd capacity.

Offline sdog

  • Devotee
  • *
  • Posts: 2039
Re: Passenger capacity in parentheses?
« Reply #12 on: November 28, 2010, 07:21:55 AM »
Quote
They are loading one after another. But next bus is waiting near and passengers see it.
That's exactly the what i meant. One bus loads/unloads the next is right behind. The other buses are not loading at the same time?

I'm mostly asking out of curiosity, in game this is actually not a big issue, and for long distance trains it can be addressed easily in the pak-set.


on another note, i always thought Kazan would be much further in the south and east! Only 300 km to Nijni Novgorod. And it looks quite fascinating on google maps.
« Last Edit: November 28, 2010, 07:35:56 AM by sdog »

Offline inkelyad

  • Devotees (Inactive)
  • *
  • Posts: 482
  • Languages: EN, RU
Re: Passenger capacity in parentheses?
« Reply #13 on: November 28, 2010, 08:06:03 AM »
That's exactly the what i meant. One bus loads/unloads the next is right behind. The other buses are not loading at the same time?
Only one. And simutrans should do it too. It is how loading queue should work.

Quote
And it looks quite fascinating on google maps.
Yandex.maps have StreetView mode.
« Last Edit: November 28, 2010, 08:11:30 AM by inkelyad »

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18422
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Passenger capacity in parentheses?
« Reply #14 on: November 28, 2010, 11:34:29 AM »
I am not sure about the passenger readiness delay algorithm - that strikes me as potentially complicated to code for what is likely to be a marginal advantage (in only a very few cases will the 'bus be sitting in the station at the same time as a large train arrives). It might also be confusing to a player not to understand where the passengers who have just arrived at the station and got off the train have gone, and why only slowly are the numbers of passengers in the station are increasing.

I think that the issue with rail passengers offloading onto 'buses is better solved by the design of 'bus routes in towns rather than by this mechanism. The best 'bus route designs in Experimental are ones that have a large number of different lines, each with direct routes to specific parts of the towns, rather than one or two meandering or circular lines that take in most of the town (the latter of which tends to be more efficient in Standard). The reason for this is that, with direct routes, passengers get where they want to go faster; however, it also helps to solve this problem, as, when a train arrives at the station, not all passengers want to get on the same 'bus.

As to comfort, incidentally, do note that comfort is not used in routing: only revenue calculation.

Offline sdog

  • Devotee
  • *
  • Posts: 2039
Re: Passenger capacity in parentheses?
« Reply #15 on: November 28, 2010, 07:48:10 PM »
Quote
...complicated to code for what is likely to be a marginal advantage.
certainly, unlikely it is worth to implement. It's a slightly complicated solution for a problem not necessary to solve. (It'd require also quite a lot of research what the real distribution of delays is, Gaussian was just a guess)

Quote
It might also be confusing to a player not to understand where the passengers who have just arrived at the station and got off the train have gone, and why only slowly are the numbers of passengers in the station are increasing.
As long as pax are 'moving' in the station, display a message at the station screen "passengers transfering"

Quote
I think that the issue with rail passengers offloading onto 'buses is better solved by the design of 'bus routes in towns rather than by this mechanism. The best 'bus route designs in Experimental are ones that have a large number of different lines, each with direct routes to specific parts of the towns, rather than one or two meandering or circular lines that take in most of the town (the latter of which tends to be more efficient in Standard). The reason for this is that, with direct routes, passengers get where they want to go faster; however, it also helps to solve this problem, as, when a train arrives at the station, not all passengers want to get on the same 'bus.
I had by far the best results with suburban rail ring lines, distributing the passengers to bus/horse carriage lines. Point to poiny lines cause too high congestion.

Ring lines don't work at all in standard simutrans, they require the "reverse line" patch. With two counterrotating lines one always gets overcrowded.


Quote
As to comfort, incidentally, do note that comfort is not used in routing: only revenue calculation.
I was reffering to this line:
Quote
Introducing different classes would, [...], require an entirely new layer of simulation of affordability, [...]
The classes, and we want to implement it as comfort, affects afford-ability, in the model on which the passenger generation is based on. However the model is so crude, that the effects can easily be neglected. One thing that could be done is logging the average comfort per pax*km and base a passenger reducing factor on this. The same way you did for car ownership. But as i said before, the model is so crude, this would be, i think, a waste of time and needlessly increase complexity.

An implementation of classes on the pak-set level is easy, straight forward and should be quite good enough.

Offline ercsim

  • *
  • Posts: 32
Re: Passenger capacity in parentheses?
« Reply #16 on: May 05, 2011, 06:49:34 PM »
I think this problem need to be revisited.

As previously suggested, there should be an option in each of the lines to allow/disallow overcrowding. With this option, overcrowding can be disallowed on long hauls, such as intercity routes.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18422
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Passenger capacity in parentheses?
« Reply #17 on: May 05, 2011, 07:43:21 PM »
Ercsim,

I don't think that that would be realistic, as, in reality, if a train with vehicles that have room for people standing are used on a long-distance route and there are not enough seats for everyone who wants to get on, people will stand. There is no practical way of stopping people getting on to vehicles whose normal seating capacity has been exceeded in most cases: it is enough that it is possible for pakset authors to specify that certain sorts of vehicles (such as aircraft, for example) cannot be overcrowded.