Problem with infection number

CityScope / CS_Proxymix

ABM framework to study spatial configuration and human behaviour at the building scale

10 stars 7 forks source link

Problem with infection number #169

Closed agrignard closed 4 years ago

agrignard commented 4 years ago

Just run Episode 1 A and B with the same scenarios. The number of infection sis very very different now.

We (@tnguyenh and I) are now starting the validation process of the model and the starting point is to identify the type of infection (direct, object, air) and be able to easily see the impact of intervention. The problem is that now 2 potentially similar simulation gives a very different output in term of infection.

@ptaillandier do you think you can have a look at it?

agrignard commented 4 years ago

Ok I have just notice that the queing can have a big effect on the replicability of the sim. For now without queing we have a more reproductible number

WIP...

agrignard commented 4 years ago

Many change has been made this last commit (for isntance taking in account the mask for air and object), so I close this issue now