David Hensle

I attempted a run with TransLink's model on our own 24 cpu, 250 GB RAM machine using the current ActivitySim "main" branch and got the following results: * Successful chunk...

ActivitySim is moving towards [explicit chunking](https://activitysim.github.io/activitysim/v1.3.1/users-guide/performance/chunking.html#explicit-chunking) instead of chunk training.

As mentioned above, I tested with the current main branch of the code and the sandag-abm3-example. The results were very similar. I ran with 100k households in chunk_training mode without...

The ActivitySim platform is supporting [explicit chunking](https://activitysim.github.io/activitysim/v1.3.1/users-guide/performance/chunking.html#explicit-chunking) instead.

Fix in the BayDAG work [here](https://github.com/SANDAG/activitysim/commit/627e2867b904633c47007e912f69b65c84c7f549) solved the issue. PR #836 is the analogous commit back to ActivitySim main repo.

I think the issue is that the original expression itself was not correct. The original implementation found the max MicroAccessTime by TAZ and then reindexed on the alternatives. However, the...

Ahh, I see. I missed the aggregation, as you point out. I think the simple fix here is to just have separate preprocessors for the sample part and the simulate...