Thoughts on shunting, adding&removing cars, etc...

[prissi]

The system with mixed long distance trains can be done. Just instead of shunting, to load unlaod a way stations in between. By the way, those long distance trains most go indeed long distance. Michigan/Detroit (or whatever industrial areas are still left) to west coasts. In between are way too little destination to require lot of shunting.

By the way, the transib railway does shunting with passenger trains. Every 500 km or so another 3-10 car were added or removed from trains, including mail cars (our had between 1 and 4 on different legs of the journey).

[Fabio]

Obviously there's plenty of workarounds.
IMHO the problem is that presently freight ops require too many different lines for my taste. A more flexible system could allow less and more sophisticated lines.

[Roads]

@Ters:

Having this happen in depots would ruin the fun in it for me. I want to see it. The trains would also need some synchronizing mechanism. That is perhaps the biggest hurdle when it comes to more complex train behaviour: the lack of a fixed predictable schedule.

Although I did imply that is what could be done, I did not mean that depots should take on the job of switching.  My thinking was simply that all the functionality for switching cars (wagons) and/or engines is there and they can be stored there.  Not only that but the routes are there as well.  Not being familiar with the code, it still seems reasonable that much of that code could simply be copied to a new entity with a new graphic and called a switch yard.

[isidoro]

On the contrary, I think that this is a nice opportunity to give another function to depots.  Whenever you want to change the consist in you route, you click on a depot.  The change will happen there and only if there are cars of that type in the depot.  Otherwise, the train will wait.

That would allow easy merging and it is naturally realistic since cars are not created from nothing.

It is also (nearly) compatible with present behavior.  And optional.

[Ters]

I just fear that it won't be as flexible if you have a fixed recording of which wagon should go where. If we imagine a train goes from A to B, where it splits into a train to C and a train to D. Some times, there might be no cargo from A to C, other times no cargo from A to D, other times something in between. In real life, this is no problem (apart from economy). It is also no problem with the current way such has to be done in Simutrans (the A-B will run continuously, while B-C and B-D will only run when full.) But if the number of wagons for C and D is fixed already in the A-B leg, because of the recorded splitting, it becomes unrealistically difficult to ensure fully loaded trains.

Or am I misunderstanding again?

[sdog]

Perhaps some bus system would be more appropriate for Wlindley's example.

New structure shunting yard, built like a station. Trains leave at fixed scheduled intervals on a route.

Railway vehicles arrrive with other trains and are removed from their consist and added to the waiting bus-consist.

At the scheduled time the bus-consist leaves, it will call at new shunting-yards, when doing so normal route-finding decides if a railway vehicle is to leave the consist.


This would require to have another level in the hierarchy, between cargo and consist.
A line would have convoys attached; Convoys have groups of vehicles attachached; groups of vehicles have a capacity and cargo.

This possibly allows also to use containers in the game. (which would be much more imortant for a good freight simulation than the shunting)

It does not work for train ferries, or piggy back, or diesel pulling electric engine, as my suggestion doesn't allow engines in the groups of vehicles.

[Roads]

Regardless of whether you expand the functionality of the depot or create a separate switch yard, the player would have to buy the cars.  If the switch yard were separate, the player would have to buy cars at the depot and move those to the switch yard to build up his inventory of various type cars - no more unlimited supply of magically appearing cars.  This would be an expense just like laying track, etc.

I had assumed that only engines were programmed to stop at various locations, the cars simply followed wherever the engine went.  I really can see no reason for tracking where the cars go, simply read what is on the consist at a destination.

When I do this manually in a game, I just click the cars off the train in the depot and click on the ones I need which I have previously bought.  Then I change the route of the train.  If however you implement switching, instead of changing the route you just add another stop, which is the switch yard and there you tell the train what its new consist is going to be provided you have the necessary cars available. From then on, the train stops at the switch yard, switches cars and continues on its way.

[rsdworker]

well splitting and joining - could be good idea even i been on alot of trains in UK - most them can spilt or join like one seen on video

Passgeners can have loco change example - my route has no power between cities - even i went on Dual mode train in demark which changes from electric to diesel

in UK we have some passengers train loco changes - the Sleeper trains commonly change locos when they enter or exit from non powered lines but mostly they stay on electric mode

so my suggestion is have speical button - SPLIT here or Join here and tell which car want split up - click on Car eg Loco and add separate order

example: Route UK to Germany using two locos

1: player adds train in depot and sets the route with Split or join points and places locos in yards and gives order - eg wait for train to arrive at station - giving train name or number

2: player adds Loco orders when split - example go to depot and wait for other train to enter

3: player starts train - the train shunts in london station with shunter and shunter detaches and returns to shunter yard for new train
the train has front loco for all way to halfway (east france)

4: once train arrives at halfway - the Loco detachs and changes with new loco

5: once train arrives at end point - with Loco at end so shunter brings new loco to rear and detach front loco

for non Loco trains - they can attach or detach from other unit but can't attach to different type eg A stock can't couple to S stock unless has loco or barrier wagon in middle to tow different type of train

[ӔO]

Here is my idea, by the way.
Each freight wagon "unit" or convoy has the usual schedule to go from A to B. Except if they don't have their own power unit, they will be carried around with locomotives that have zones. If they have their own power they retain their normal simutrans behaviour.


Convoys with locomotives, or locomotives only, can have a "continuous zone" they can be designated with. In this case, it is a small shunting locomotive.


The shunting locomotive can designated pickup and drop off yards.
In this case, the only thing this shunting locomotive does is pickup full units from 3 factories and drop off in the yellow yard.


Once the factory yards are empty, the shunting locomotive will pickup any other empty units, serving the same line, in the yellow yard and drop them off at the factories


The same zone designation should give the same results for a mainline locomotive. The mainline locomotive should have a minimum and maximum length allowed for the convoy at its designated yards. In this case, it's a maximum of 10 wagons total, although for simutrans it should be in tiles. Units are automatically assembled once the locomotive is ready.


The mainline locomotive dropping off the units at its destination.


Units automatically disassembled and delivered to various factories by shunting locomotive.
Mainline locomotive picks up empty units to be brought back to the initial factories.



---
edit:

There also could be a 'holding area' for the locomotives if they are not doing anything. They should only head off to their yard or station if there's something to pickup, otherwise I think they can occupy any empty space at a yard or station. The only caveat to not having a holding area is what to do if an idle locomotive is blocking another train trying to enter the yard.

[Roads]

AEO, it seems like you have put a huge amount of thought and work into this but I just don't get the reason why wagons should have routes or tracked in any way.  To me only engines should have routes, wagons simply follow wherever the engine takes them.

[ӔO]

well splitting and joining - could be good idea even i been on alot of trains in UK - most them can spilt or join like one seen on video

so my suggestion is have speical button - SPLIT here or Join here and tell which car want split up - click on Car eg Loco and add separate order

I think this is the caveat with my idea. For passenger multiple unit trains, such a function would be necessary, since all convoys have their own power.  Something like a 'Y' line or short turn combined with long distance line seems simple, but multiple joining and splitting between various multiple units can become quite tedious if the other line cannot be specifically defined.

AEO, it seems like you have put a huge amount of thought and work into this but I just don't get the reason why wagons should have routes or tracked in any way.  To me only engines should have routes, wagons simply follow wherever the engine takes them.

Ah, that's because I'm used to the north american system. Quite a lot of factories have their own drop off and pickup yard that a shunter will move around. I think this method would also help simplify the complexity of freight schedules themselves. It's currently impossible to have a single locomotive serve two or more branch lines, so this would allow that to happen.

Also to retain old behaviour, which might be desirable for block trains

[greenling]

Aeo
that what you here present have was until 1986 in germany by longdistanstrain in use.

Here is my idea, by the way.
Each freight wagon "unit" or convoy has the usual schedule to go from A to B. Except if they don't have their own power unit, they will be carried around with locomotives that have zones. If they have their own power they retain their normal simutrans behaviour.


Convoys with locomotives, or locomotives only, can have a "continuous zone" they can be designated with. In this case, it is a small shunting locomotive.


The shunting locomotive can designated pickup and drop off yards.
In this case, the only thing this shunting locomotive does is pickup full units from 3 factories and drop off in the yellow yard.


Once the factory yards are empty, the shunting locomotive will pickup any other empty units, serving the same line, in the yellow yard and drop them off at the factories


The same zone designation should give the same results for a mainline locomotive. The mainline locomotive should have a minimum and maximum length allowed for the convoy at its designated yards. In this case, it's a maximum of 10 wagons total, although for simutrans it should be in tiles. Units are automatically assembled once the locomotive is ready.


The mainline locomotive dropping off the units at its destination.


Units automatically disassembled and delivered to various factories by shunting locomotive.
Mainline locomotive picks up empty units to be brought back to the initial factories.

[Ters]

AEO, it seems like you have put a huge amount of thought and work into this but I just don't get the reason why wagons should have routes or tracked in any way.  To me only engines should have routes, wagons simply follow wherever the engine takes them.

I think a wagon needs to know where it's going. How else is the game supposed to know which wagon should be dropped off / sent where on the empty return journey? Isn't that also how it works in real life?

[Roads]

Guys, guys, guys! 


I realized after I previously replied to AEO's post that we are talking about entirely different concepts here.  It looks like more than two and I'm going to attempt to break them down:


1.  Giving the player an automated way to change the consist.  This does not include displaying the process or building any kind of infrastructure for switching.  The reason you don't have to track the wagons is that the computer does not have to know where they are going or what they are carrying until it is delivered.  Unless my guessing is way off, this is exactly how it works now.  When I create a new route and add cars to it, say to pick up coal and then send it to a cattle ranch, I don't get an error saying I have the wrong cars in my consist.  The train simply does not pick up anything.  This IMO as it should be.


Ters, how the wagons know where to be dropped off is what you set in the consist at any particular stop.  At the depot now, the wagons know where they are going because it is part of the route, it is exactly the same idea.


2.  If I understand AEO, his idea is to make the actual switching process part of the game.  I haven't given this much thought but at first blush I don't care much for the idea.  To me, once the process is understood it would simply be another task to do requiring very little, if any thought, much like laying track.  AEO has already had the fun figuring out how to do it.


3.  This idea of splitting and joining passenger cars is not the same as simply changing the consist, it is a new idea as such would require much more in the way of programming.  I'm simply bowing out of that discussion as I'm not really into carrying passengers anyway.


 

[ӔO]

@roads Yes, there are indeed different ways of doing this, depending on what is desired.

Right now, I'm having a slightly difficult time trying to wrap my head around how #1 works.
I'm guessing you mean the game only 'knows' where the goods want to go and if there is a route that carries those goods, then that's what the game does. Yes, I do believe that's how the game currently works.

My idea is basically that, except instead of the goods being transferred at railyards, those goods are packaged into containers and delivered directly to the receiver. It's just that there needs to be a few locomotives to complete the route, instead of all of them being fully powered convoys with their own wagons.


I just don't quite understand how one can automatically change consists -I guess like RT3 behaviour?- without having a bunch of empty wagons piling up or being magically transported back to their origin.

my idea should take care of empty and full wagons piling up, to an extent, because it's still up to the player to manage their system so that it's not overcrowded or undersized. This is where it is handy to retain the classical ST behaviour with block trains, so that players have a choice.


Theoretically, the same system can be implemented with semi-trucks.

oh, yeah, one thing I forgot to add... There also should be a holding area for locomotives at the yards, but I'm not 100% sure of how it should be implemented.

[Ters]

I don't quite follow #1 either. Are you suggesting that a locomotive essentially automatically changes between two (or more) different lines by putting both lines together in it's schedule, along with an automated change of wagons? If so, then in my mind, this would require a lot of coding for very little gain (some eye candy and saving the purchase cost of a second locomotive), though the Japanese might have done some of it already. There is also the problem that the two lines/consist runs must be synchronized. By that, I mean that the locomotive must switch between the two consists in a regular pattern. With my lines, such a pattern would need to be irregular.

This would however fit nicely into my idea of a scripted schedule, but then you'd need to have some scripting skills to use it.

[Roads]

@AEO

I never played RT3 so not sure what you are talking about there.  My idea is simply to automate that which can be done manually now by sending a train to the depot, removing cars and adding different cars then sending the train on its way.

You are right.  This method would absolutely allow for cars "piling up" at switch yards if the player did not include in his planning of the route how to handle the cars that were dropped off at the switch yard.  This would have to be part of the thinking "how to" when setting up a route.

I need to study what you have done more.  I'm afraid I was too dismissive, having already in my mind what would work.

[Ters]

In RT3 you tell a train to go from A to B with a certain set of cars, from B to C with possibly another set of cars, and so on. Cars magically appear when needed and disappear when not needed. I don't remember how the wait for load worked, though. And that is something that might be a challenge to get right when different cars come and go. It's a problem already today with mixed trains.

[Roads]

@Ters


I am going to try to give you an example of what I'm talking about:
Say you have in fairly close proximity a steel mill and a sand plant.  To the east is a cement mixer and a cotton farm, to the far north is a textile mill and somewhere between  is shopping mall.  To the south of the cement mixer and cotton farm is an ore mine which is not a long distance to the south.


Here is what I would do.  Pick up the sand (bulk) and carry it to the cement mixer.  Stop by the depot (switch yard) drop the bulk cars and put on crate cars.  I would pick up the cotton and carry it north to the textile mill, pick up textiles (same consist) and drop them off at the shopping center.  I would then travel to (dead head here) the switch yard where the bulk cars are stored (remember this trip wouldn't be that far away) pick up the bulk cars there and dead head south to the ore mine.  There I would pick up the ore and deliver it to the steel mill which is near the sand plant.


You are making some dead head runs here but not as much as if you created individual routes for these trips.  And to me that's the key - not every trip has to be loaded - you just look to see what is most cost effective for whatever situation you have.


Modify:  No doubt there could be much better examples of the things that can be accomplished because I kinda hurriedly thought this one out.

[Ters]

There are a few problems here, given the way Simutrans works otherwise (unless just in time is off). There might not be textiles waiting at the textile mill when you've dropped off cotton. It is also possible that there are no cotton heading for the textile mill, though there are textiles there to pick up. One might therefore still end up with only half a load on average back and forth. And with more trips mixed in, it just gets worse.

[Fifty]

I like AEO's idea of having factories fill cars or containers instead of full convoys. (This idea could even lend itself to economic simulation of transloading costs of moving cargo from railcar to truck and the benefits of containers, but let's not get ahead of ourselves). However, it could likely be even simpler if you maintain the current scheduling design instead of zones, and merely change wait for load. Instead of waiting for a convoy to be full, a locomotive (or set of locomotives) would wait until it had a certain tonneage bound for its destinations, or a certain number of tiles, whichever comes first.

For example, one industry that is somewhat difficult to supply as in real life is the construction wholesaler: you need three different trains for three different cargos to deal with differing railcar sizes and factory demands. With this system, you could set up a single locomotive from a central yard, tell it to wait for, say, 400 tons of goods or a 5-tile train, and then it would attach wood, steel, and concrete cars as they are needed and produced and bring them to the wholesaler. The same could work for trains between yards.

[Roads]

@Ters
That's why you have the wait ability and also load percentage ability.  If most of the time you are carrying goods when the train is moving, it will be profitable even without a full load.

Problems?  Yes of course.  That's what makes it a game, you simply need options to deal with those problems.

@Fifty
If I understand correctly what you are saying, yes I absolutely would love to see something like number of tiles full or number of cars, etc.  The percentage works fine for a consist of same type cars but does not work for a mixed consist.

I still have not yet thought through AEO's suggestion.

[Sarlock]

Been thinking about this a lot... here's my thoughts:

First, in Roads' example, you're not saving any time at all.  In the end, all empty containers return to their point of supply and then travel loaded.  The locomotive is doing a lot of flipping of cars but you're not actually saving any transit time than under another scenario.  If you ran your train with a consist of a few of each type of car that you need, balanced for load requirements, you'll achieve the exact same result, same load amounts per locomotive and same amount of dead head trips without all of the switching.

Where switching is commonly used in real life is in industrial areas to gather up materials from various industries and aggregate them for a longer voyage.  I'm not sure how it works in Europe as the travel distances are different, but where goods may travel thousands of miles, this is the method generally used.  Industrial spur lines gather the goods and aggregate them in a switching yard where they are put together in to a longer train for the journey.  At various points along the trip, the cars are removed and delivered to their end users and new cars are placed on to the main train and it carries on its long voyage.

From what I can see, though, we can already mostly simulate that in Simutrans.  Intermodal yards are possible by having trucks pick up goods from train stations... that's easy to do.  Transferring the rest of the material across a long freight line is just as easy by carefully constructing your lines and having your main freight trains have the right balance of train cars to carry all of the freight that goes down this line.  More thoughts later, out of time for now.

[Ters]

Where switching is commonly used in real life is in industrial areas to gather up materials from various industries and aggregate them for a longer voyage.  I'm not sure how it works in Europe as the travel distances are different

In Norway, tracks to individual industries are all but gone. The few that are left are, as far as I know with just one exception, served by dedicated trains. I think these trains are usually just 10 to 20 cars long (400 meters is considered a long train in Norway). For lesser quantities, rail has lost to road. The road vehicles might transfer to rail at a terminal, or go all the way even on multi-day trips. But until late in the 20th century, mixed freight trains could be assembled in almost any town. Usually at the single station, as there were and are few dedicated freight stations in Norway.

[ӔO]

From what I can see, though, we can already mostly simulate that in Simutrans.  Intermodal yards are possible by having trucks pick up goods from train stations... that's easy to do.  Transferring the rest of the material across a long freight line is just as easy by carefully constructing your lines and having your main freight trains have the right balance of train cars to carry all of the freight that goes down this line.  More thoughts later, out of time for now.

One problem with the current implementation, is that you would need very long yards for very long trains, where as in real life, a long train is broken down at the yard. Hopefully, with what I've suggested, long trains can be assembled and broken down as long as there are enough empty tiles at the yard.

[Sarlock]

In Norway, tracks to individual industries are all but gone. The few that are left are, as far as I know with just one exception, served by dedicated trains. I think these trains are usually just 10 to 20 cars long (400 meters is considered a long train in Norway). For lesser quantities, rail has lost to road. The road vehicles might transfer to rail at a terminal, or go all the way even on multi-day trips. But until late in the 20th century, mixed freight trains could be assembled in almost any town. Usually at the single station, as there were and are few dedicated freight stations in Norway.

It's similar here, a lot of industrial trunk lines are gone and road freight has replaced that service.  Road is far more versatile/flexible and for short journeys is very competitive on a $/km ratio.  Rail dominates in longer journies and where turnaround time is not a significant factor in cost (since rail transport is slower point to point).  And I bet that a lot of the industries that still have freight delivery by train would probably just be built with truck freight if they were built new today.  Almost all rail freight is point-to-point bulk deliveries (coal, oil/fuel, minerals, etc) or intermodal these days.

As AEO mentions, the limitation of 1 station tile per 2 cars for stations creates very long freight stations (12 tiles long for a full length train) which is a bit of a pain... what might work is the ability to load a freight train with a smaller station at the expense of a longer loading time (24 car train loading at a 6 tile station would take 2x as long to load).  Passengers would still need a proper ratio of stations to cars.

[Roads]

@Sarlock,

Time was not a consideration in my scenario.  If time is to be the all important factor in Simutrans, both for passengers and frieght, then of course one way loads are likely the only way to achieve that.  Also if that is the case, I'm done here and off to work on my game.

[Sarlock]

I was only referencing time with respect to how things are generally done these days from an industrial freight standpoint.  When time is important, freight options lean towards truck... for large bulk items where time isn't important, rail is more likely to be chosen (and if rail is available, of course).

I just ran some freight train scenarios that I had only thought of in my head but hadn't put in to implementation yet (I was planning to do so in my current game, just haven't done it yet).

I ran a trunk link between two urban areas with a 12 tile freight station at each end.  Then I ran spur lines off of the main station to various industries in the area.

Industries A,B,C,D ---> Western Freight Station ==========MAIN TRUNK LINE========== Eastern Freight Station ---> Industries E,F,G,H

Attached to the main stations are small 2-3 tile small stations that serve the spur lines that serve industries in the area.  Small, cheap locomotives are used to pull 3-4 cars to each of these stations and gather up materials and deliver from the main station it serves.  When a single locomotive serves several industries, it can have a mix of cars appropriate for the industries it serves.  These lines were all profitable even when running empty half the time (freight car costs are balanced with this in mind).

Then the main trunk line has a consist of cars that includes every freight type used in industries A-H.  3 locomotives and 21 cars of a mix that I balanced to keep the stations from stockpiling materials.  The main line serves to transfer materials back and forth down the trunk line to feed the spur line industries.  Trains were very profitable doing this run.  Often cars were full both ways with different cargo types that used the same car (ie: iron ore one way, coal the other).  The more industries you can serve with your spur lines, the more full your trains will be going in each direction.  A lot of freight shares the same car types.

Nothing new here, I'm sure many people have been doing this for ages, but I hadn't thought of it from a larger vision like this before... this same philosophy could be employed on a much longer freight main line crossing a country, for instance, stopping at a few key main freight stations along the way.  Everything else would then be served by spur lines and truck freight.  It looks really cool, works, is profitable, and accomplishes the real life simulation quite nicely.  The only thing you don't get is the shunting of actual cars but that's where imagination kicks in   I was always much more point-to-point oriented with my freight lines.

One difficulty that I had was when the same car would be required to transport different materials, ie: coal and iron to a steel mill.  I had a bit of trouble ensuring that I'd get the right mix of coal/iron to feed the steel mill.  When the spur line train reached the main station and there was ample coal and iron stockpiled, it would be nice to ensure that I got the right amount of each to fill the cars rather than overload with something that I didn't want.  In pak128 coal/iron requirement for a steel mill is 1:4 coal:iron, but the train would often load up with coal when the steel mill was already loaded with coal and needed iron to produce more steel.  I imagine that once the steel mill completely filled with coal it would then choose iron more often.  On the flip side, my job is to run a transport company, not optimize industrial production, so it's up to the factories to balance themselves, I guess.... but it also does mean that I don't get any steel for the return trip when I deliver coal instead of iron.

[ӔO]

oh, if anyone wants, here is the sketchup file I've made to illustrate my ideas.

It should make drawing out diagrams easier.

[hreintke]

LS,

Like AEO I also am thinking how to use a "container" concept for transporting.

Would it be possible to add a container as a input/output material for factories ?
In that way, if containers are available at a factory and trucks/trains &lines capable of transporting are in reach, the factory can package the goods into containers and transport to destination will happen with the current transport mechnisms is place.

If Steelmills can be configured to create containers, they will go from there spreading across factories filling them.
At the destionation, when emptying a container, the factory either can fill with own output goods or the empty containers can be shipped to factories needing them.

There is also the need to destruct containers after (long) usage (in capyards ?) otherwise there are more and more containers in the game.

The transport vehicles for transporting the containers (truck/train/ship) would be very much "multi-purpose" and hopefully be able to run fully loaded in both ways.

Herman

[kierongreen]

Containers are already used by some paks for certain goods. It is up to pakset maintainers to decide how many categories of goods there are - if a pak just had one goods category then different images could be specified for each good type and you'd have behaviour similar to RRT (except that you would always have a fixed number of waggons).

[Sarlock]

Everything that is a container product could be coded to goods type "goods"... you could recompile the pakset for just the factories in question and recode the goods types accordingly.  Of course, this doesn't give you a historical reference as container intermodal transport is a more recent development in history.  It also means that you could potentially deliver odd goods to end user factories, like delivering steel to a supermarket...

A question on game mechanics: how does a train (truck, ship, plane) prioritize which cargo it picks up first?  ie: if there is 500 widgets waiting for one factory and 500 widgets for another that can use that line and the train can only hold 500, how does it decide which of those 500 it loads?

[Combuijs]

A question on game mechanics: how does a train (truck, ship, plane) prioritize which cargo it picks up first?  ie: if there is 500 widgets waiting for one factory and 500 widgets for another that can use that line and the train can only hold 500, how does it decide which of those 500 it loads?

Everything for the first station it encounters is loaded, then for the second station etc.

[Sarlock]

I should have rephrased that: the deliveries go to the same station.  ie: You have iron ore and coal at station A and it gets delivered to station B where the steel mill is.  Station A has lots of iron and coal available: how does the train select what to put in to its cars?

[Roads]

This is going to sound cynical but from my experience, I can tell you how it priortizes.  It picks up what you need the least.  That is not what happens initially.  The first trips everything goes swimmingly.  Just when you start patting yourself on the back thinking what a great job you've done, then you begin to see things like 20% of what you need, 80% of what you do not need.  I have not let this situation run its course to see what happens when the receiving factory gets too much of one item but I expect the train would just set there until the factory used enough of the raw material in question to unload the train.


Modify:  What I normally do in a situation like this is to build another station with coverage of only the raw material I needed and route my train to there.