Dear Frank, Christophe, and all,
I'm sure many of us on Simsoc will take an interest in viewing such
models when you're ready, thanks!
In the meantime, there is plenty for Social Simulation modellers to discuss.
As some of you may know, the UK government has (at time of writing!)
adopted a rather different approach to other European countries. In
particular, most schools and workplaces remain open here (unless closed
by more local decisions). I know enough about disease modelling to know
not to rush into judging others' conclusions until I've seen the
working... (But I *am* looking forward to them publishing more details
about their modelling.)
However, from the comments of the UK's Chief Scientific Advisor and
Chief Medical Officer last week, two modelling lessons have struck me in
particular.
Modellers' lesson one is the importance, when designing a model, of
continuing the story.
For example, it would be easy for a modeller to think thus:
1) Children go to school, cough over each other, spread virus.
Therefore, close schools. Disease thwarted!
But a wise modelling team know they have now to continue the children's
stories...
2) Children stay at home. Under UK law, that means small children must
be looked after at home by someone older.
a) If that person is a parent, there are consequences for that parent's
work. If that parent works in healthcare, there are consequences for the
UK's ability to deal with the virus.
b) If that person is a grandparent, then the child is now in contact
with - and able to cough over - a member of a group much more vulnerable
to the virus.
Another way to continue the story is:
3) One parent agrees to host play dates for other families' children, to
stop them getting bored. Thus the children are not in social isolation
(though their social contacts are at least smaller than at school).
So what's the best advice? To close, or not to close, the schools?
Modelling lesson two, I would suggest, is to reply that we don't know
the answer in advance.
We'll have to estimate, as best we can, the adoption rates and infection
rates attached to these social practices. And then run the simulation
models. And until then, we don't know what numbers are going to come out
of them. Unanticipated outcomes from social systems is our business. The
sizes of some of the effects of particular policies may yet surprise us.
E.g. most of the success in delaying infection may turn out to be coming
from increased hand washing, or self-isolating when you first display
the symptoms, rather than from school closures. (The official advisors'
comments suggest to me that this is what their models have been saying.)
Thus, while the UK government's advice has already prompted at least two
critical joint letters (one mostly from behavioural scientists, the
other mostly from mathematicians and complex networks people), I myself
won't be signing any such letters ... yet.
Comments and other suggestions welcome, and I look forward to everyone's
models.
Christopher Watts
On 16/03/2020 13:53, Frank Dignum wrote:
> Thanks for sharing this.
> We in UmeƄ have just (one hour ago!) started to develop a more mature
> agent model for a simulation that could give more insights in the
> effectiveness of different measures. If we get anything interesting
> out of this effort in the next week or two we will certainly share.
> The simulation will be in NetLogo in order to work quick and also
> allow other people to work on it. Of course we will share that code
> once we are confident that we have something worth to share.
> Best regards,
> Frank Dignum.
>
> On 2020-03-16 14:31, Christopher Watts wrote:
>> For those of you who haven't seen it yet, the Washington Post has
>> discovered agent-based simulations!
>>
>> https://www.washingtonpost.com/graphics/2020/world/corona-simulator/
>>
>> While these models are too simple - and the article's author too
>> careful - to make serious policy recommendations from them, it could
>> be a good opportunity to explain to colleagues / friends / family
>> what the agent-based modelling approach can be used for...
>>
>> Keep healthy,
>>
>> Christopher Watts
>>
>>
>
--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
########################################################################
To unsubscribe from the SIMSOC list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=SIMSOC&A=1
|