-1

Ok, what i want is to create cars which havn't predefined flows or trips, because i want to make some cars circulating randomely on a closed network.

So i have a closed network and the only way i found to make something kinda similar to what i want is to generate a lot of differents trips with the script "randomTrips.py" in the SUMO/tools ...

But in fact this isn't a chaotic simulation because it's reproducible and that's not what i want ! I really want that each time an agent comes across an intersection he makes a random choice of where he'll go.

So if someone got an idea I apreciate

Xeyes
  • 583
  • 5
  • 25

2 Answers2

0

Huu nobody got an idea ?

This environement will be used for deep Q learning application and i realy need my agents to randomize their choices

Xeyes
  • 583
  • 5
  • 25
0

If you don't need the agents to start at the same time and/or place in every run you can just rerun randomTrips.py before every run. It will output different trips each time. You can also include a number of random intermediate waypoints using the "-i" option. If this is still not enough you could have a look at rerouters and place one of those on every possible edge.

Michael
  • 3,510
  • 1
  • 11
  • 23