Trains Maintain Their Lanes (hey that rhymes)

[andreharv]

The code may already be in here (just needs to be reorganized) but as quad tracks are soon to become a thing, it would be nice to have vehicles (trains in this case) favor the lane they are on rather than switching lanes for a slight distance difference in curves. The exception to this, of course, would be if the placed stops mandated a lane change which would mainly be an issue for quad tracks with central bypass lanes. Otherwise there would be no reason for a lane change as all junctions naturally connect all compatible lanes. I don't remember if the pathfinding is a weighted system but basically this would all be fixed if more 'weight' was added to changing lanes. That way trains would not do it unless there was no other way to reach their endpoint or the speed difference between the lanes outweighed the lane change (which I would imagine would be rare).

[andreharv]

Ok giving this 2 seconds of thought, it is very similar to #89 Medians. It will operate the same as a road that has medians separating each lane, except at 3+ network junctions.

[originalfoo]

The issue of vehicles changing lanes for a slightly shorter distance at curves also affects highways in a most annoying manner: #112

[VictorPhilipp]

Path-finding uses average segment length for cost calculation, not lane length. Your issue must be caused by something else.

[VictorPhilipp]

@andreharv This only occurs for one-way tracks, correct?

[andreharv]

Yes. This is only an issue on one way tracks and quad tracks (MOM). MOM QuadTracks Build (it's near complete). You can use it to test pathfinding if that helps. I'm sure that if anyone comes out with quad tracks for passenger rail the same logic will apply. It will be a lot easier to see when putting down quad tracks and stations. All the lanes are the same speed except for the bypass station tracks (outer tracks are stations and inner tracks are bypass).