Started by Sirius, April 10, 2020, 09:12:10 AM
0 Members and 1 Guest are viewing this topic.
Quote from: Freahk on May 11, 2020, 09:32:39 PMKeep in mind, we will need to ensure that congestion data does not change within our calculation timeframes.Using old congestion data should work well. We already maintain such "stats" for the current and the previous month.Maintaining a third, more recent but not live congestion "stat" should not be an issue.
Quote from: Freahk on May 14, 2020, 11:05:14 AMSo the weight in between two koors is kind of randomised.Hopefully it is modelled in a way that relates multiple nodes to a koord, otherwise Dijkstra itself would break.At least from the descriptions it sounds like it's relating multiple nodes to the same koord.I did not study route search code in simutrans so far, but I guess I should do this just now...