Let me respond to Nick who knows the land use models that I refer to here well. Basically the problem is often that our aggregate models where the focus of interest is on the location of aggregates of population and other activities in zones – like census tracts, blocks, output areas and so on – sometimes need to be made consistent with models at a finer scale, particularly transportation. The key issue is how can an aggregate model that we can validate in the usual way for we have detailed data on the aggregates be made consistent with an agent-based model where every household in the aggregate is an agent. The way we have thought about this – we have produced demonstrations – is to define the land use model at the smallest spatial scale we can to make sure that we meet the statistical criterion (of large numbers) that enable us to fit the model using maximum likelihood whatever. Then having got a ‘good fit’, we use this model as a control on the disaggregation of population to the household level – to the agent level using all sort of diverse data that helps us define the actual location of each household. Inevitably error creeps in but then our data set of households is a kind of synthetic data set which we know meets certain control totals and hopefully is sufficiently close to realism to be generally useful in calibrating the agent based model.
We had a go at doing this a few years ago which is briefly alluded to in one paragraph in the attached paper. We are also beginning to think about this again with respect to linking our aggregate land use models to ever more detailed transportation models that operate at the level of households and cars and vehicles like in MATSIM et al.
Mike
______________________________________
On 13 Sep 2018, at 08:01, Nicolas Malleson <[log in to unmask]> wrote:
Hello again SIMSOC,
I just wanted to briefly thank the large number of people who replied to my original email. Some really interesting points have been raised; all of which have given me and my co-authors a lot to think about. I’m going to spend a few days considering and collating the responses, and will provide some kind of summary which I hope might be of use to others.
I also should just briefly add, in case it wasn’t entirely clear in the first email, that in the context of *this particular paper* I think the reviewer’s comments are perfectly reasonable. Whilst we wont be able to validate this model to the extent that some empirical models have been, I agree with the reviewers that some kind of validation is necessary.
All the best,
Nick
—
Dr Nick Malleson
Room 10.114 (Manton building)
School of Geography, University of Leeds
[log in to unmask]
http://nickmalleson.co.uk/
0113 34 35248
On 12 Sep 2018, at 08:03, Nicolas Malleson <[log in to unmask]> wrote:
Hi SIMSOC,
I was wondering if anyone has any thoughts/advice about a difficulty that I’m having with validating a model. This is in response to (very fair) comments by reviewers on a paper that is under review, so I will talk about the problem in general terms. I think the discussion should be of interest to others on the list.
Colleagues and I have built a spatially-realistic agent-based model with agents who move around. It’s based on a real urban area. We have used an a-survey to calibrate the behavioural parameters, such that the agents behave in a way that is consistent with the results of the survey. The survey is national, so not specific to our study area. We put the agents into a virtual environment, let them go, and see what happens.
The reason for creating this model in the first place is that we have no data on the spatial behaviour of the real individuals in our study area. So we’re hoping that by implement behaviour that is consistent with the results of the survey, the agents will give us some insight into the real dynamics of the case study area.
But how do we validate the model? Assume that there are no empirical data available for our study area (it is possible to try to stand on the road and talk to people, but this is probably out of scope). What should an aent-based modeller do when they have an empirical model but no empirical validation data??
All the best,
Nick
—
Dr Nick Malleson
Room 10.114 (Manton building)
School of Geography, University of Leeds
[log in to unmask]
http://nickmalleson.co.uk/
0113 34 35248
########################################################################
To unsubscribe from the SIMSOC list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=SIMSOC&A=1