faq 37683204

Do plans have to end at the same place where they start?

by Kai Nagel on 2015-10-26 18:47:59

---

when I ran it and it worked with this warning :

2015-10-25 09:29:14,366  WARN CharyparNagelActivityScoring:255 The first and the last activity do not have the same type. The correctness of the scoring function can thus not be guaranteed.

Does the trip HAVE to end at the initial/first origin?

Thanks a lot

Deema

---

Comments: 2

---

Re: Do plans have to end at the same place where they start?

by Kai Nagel on 2015-10-26 18:54:24

If the first and the last activity are of the same type, ...

... then default MATSim uses "wrap-around scoring".  That is, the last and the first activity are merged, with duration from lastActivity.startTime until firstActivity.endTime + 24h.

For this to work, it is not necessary that these activities happen at the same location.

This is necessary because of the 24h time constraint; it results in a "value of time as a resource".

If the first and the last activity are not of the same type, ...

... then this does not work: We could merge the durations, but we would not know which typical duration to use in order the compute the score.

There is some discussion around this at https://matsim.atlassian.net/browse/MATSIM-387 .

---

Re: Do plans have to end at the same place where they start?

by Johan W. Joubert on 2017-01-20 05:30:20

Deema, in South Africa I typically use different activity types for the first, last, and any other home activities. I then also give them different typical durations in the Config. Whether this is theoretically right, I don't know, but intuitively it (at least to me) makes some sense: I value being at home not by long the total hours are, but rather "how rushed am I from when I get up until I have to leave the house" (in the morning), and "how close to bed-time do I only arrive at home" (in the evening). And then home-based activities that occur throughout the day are treated separately(if there are more than two in the plan).

This is what I found the distributions to be:

For the first activity (coded h1):

For subsequent activities (coded h3):

And for the final home activity (coded h2):