Closed wese-da closed 6 years ago
is this still relevant?
Kind of... Since we want to model any German district (including some pretty big cities) sampling is still relevant i.m.o.
First problem is well known: In a 10% sample, one car is inflated to represent 10 cars which affects traffic dynamics especially on small links. Worst case: one car gets stuck in a living street and can't move forward because the capacity is too small.
Second problem is: There aren't really any systematic studies on this topic that I am aware of. The big question here is: If we sample a population, what kind of links do we need to remove at which time?
Maybe it's not first prio but I don't want to drop this issue yet.
solution for 10% sample hard coded. would be nice to retackle this issue in future e.g. bachelor thesis
meaning: e.g. 10% sample should be routed on secondary network, 100% can use a detailed one.