Some personal changes
Posted: 02 Feb 2018 15:53
I made some personal changes that I would like to share, in case some people are interested in implementing them further!
Just for the record: I branched my code of JGR's patch git repository. Obviously anybody can do with the changes what they wish, and I am sure they could be turned into a diff against the base game by someone who is more experienced in that than I am..
My changes can be found on this branch on github. I had a few issues getting the code to compile with Visual Studio, so ignore the first few commits.
The changes so far
Running early economies (1870's) you can easily need a lot of delivery vehicles as they are slow and have low capacity. This gives two issues:
1. Vehicles on a conditional delivery system (goto A, if loaded go to B, goto A, if loaded go to C, etc) will circle the pick-up station
2. Vehicles tend to pick the same drive-through loading area as the one in front of them, causing a block despite there being open bays.
So first off, I made vehicles which see their next actual order is the same station as they are at immediately go back to loading:
This has the consequence many vehicles can be waiting for their delivery station to have available things and depart immediately:
This also works for ships and trains.
I also removed a check from the refit code that erroneously seems to think vehicles are already refitted, causing them to not refit in this home station scenario.
For the second point, I modified yapf to check the tile in front of the current vehicle for other vehicles; and for every vehicle there note the tiles they are headed to right now. These tiles are then given a higher cost when traversed, and an even higher cost if they are loading areas. This means vehicles are willing to travel farther if it means not going the same way as the person in front of them. This can help clear up clogged cities, too, as vehicles are more willing to diversify.
Before:
With modifications:
This idea came about when I noticed vehicles make the wrong loading area choice because they make the decision to turn into it before the vehicle in front of them has entered the loading tile; meaning they think the tile is free whereas really it is already occupied.
This changes are not guaranteed to be stable or good, but as I made them and am enjoying my networks more with them, I thought I would share.
Just for the record: I branched my code of JGR's patch git repository. Obviously anybody can do with the changes what they wish, and I am sure they could be turned into a diff against the base game by someone who is more experienced in that than I am..
My changes can be found on this branch on github. I had a few issues getting the code to compile with Visual Studio, so ignore the first few commits.
The changes so far
Running early economies (1870's) you can easily need a lot of delivery vehicles as they are slow and have low capacity. This gives two issues:
1. Vehicles on a conditional delivery system (goto A, if loaded go to B, goto A, if loaded go to C, etc) will circle the pick-up station
2. Vehicles tend to pick the same drive-through loading area as the one in front of them, causing a block despite there being open bays.
So first off, I made vehicles which see their next actual order is the same station as they are at immediately go back to loading:
This has the consequence many vehicles can be waiting for their delivery station to have available things and depart immediately:
This also works for ships and trains.
I also removed a check from the refit code that erroneously seems to think vehicles are already refitted, causing them to not refit in this home station scenario.
For the second point, I modified yapf to check the tile in front of the current vehicle for other vehicles; and for every vehicle there note the tiles they are headed to right now. These tiles are then given a higher cost when traversed, and an even higher cost if they are loading areas. This means vehicles are willing to travel farther if it means not going the same way as the person in front of them. This can help clear up clogged cities, too, as vehicles are more willing to diversify.
Before:
With modifications:
This idea came about when I noticed vehicles make the wrong loading area choice because they make the decision to turn into it before the vehicle in front of them has entered the loading tile; meaning they think the tile is free whereas really it is already occupied.
This changes are not guaranteed to be stable or good, but as I made them and am enjoying my networks more with them, I thought I would share.