Stuck pt agents : is it usual in Matsim?

[subrina0013]

Hi.. I am trying to simulate around 3 million people and found 3% of the agents who take pt (600,000) are stuck. Is it a normal behavior in Matsim or I am doing wrong somewhere? Is there a way I can do better? Thanks

[jfbischoff]

Are they stuck after iterating for quite some time or only in the first iterations? Having some people stuck in the first iterations is rather normal, but it should level out rather fast.

Some possible reasons are:

• The schedule does not fit time wise or spatially to the rest of the data, e.g., the schedule ends at 24:00, but you are simulating later on. Or there is a mismatch in coordinate systems.
• Agents try to leave at times when there is no transit service
• Agents have their activity locations too far away from transit stops, try experimenting with the radius.

[mrieser]

Via includes a Stuck Agent Analysis to help analyze such problems. It gives an overview of when and where agents get stuck. If the agents get stuck all over the day, it's likely something with the schedule or configuration. If most agents get stuck at the end of the simulation, then the plans might just be too full for those agents or, as jbischoff mentioned, they spent too much time walking to/from transit stops.

[subrina0013]

I saw many of the agents walk long so may be that's why they miss the bus. And yes, they are mostly stuck at the end of the day. I will try to experiment with the radius. But is there anything else I can do to improve that?