 |
 |
Hi, To follow on to Oliver's note, while I have not used real data files in NetLogo, I have used the Mathematica-NetLogo link for a simulation experiment. The link allows one to control NetLogo via Mathematica, which allows for automated changes of parameters over as many runs of the model as required (and one could in theory change model parameters mid run as well). Output statistics are then saved in Mathematica, which allows for handy manipulation of the resulting data. Other than its relatively slow run speed, I was very happy with the whole set-up. I am happy to share more details off-line if anyone is interested. Cheers, Abby Postdoctoral Fellow University of Technology, Sydney School of Finance and Economics On Tue, Jun 30, 2009 at 7:39 PM, Lawson, Tony<[log in to unmask]> wrote: > Oliver Mannion wrote > > Hi there, > > We are a University based research centre embarking on the development of a > microsimulation model of health outcomes using empirical longitudinal data. > The end product will be a software program for use by external end users. > Our complete list of requirements is attached. > > > > Hello Oliver, > > I am a Ph.D student and I used NetLogo to set up a demographic microsimulation of the UK population. > > Data files at the household and individual level are read in from external files. The simulation then runs by calculating annual transition probabilities for births, deaths partnership formation and dissolution based on logistic regression equations. The model is updated each year which corresponds to one tick. It is a stochastic simulation and agents are cloned by using a simple command. A range of graphs and simple statistical functions are available. > > The simulation runs on a desktop computer and takes about one minute for each simulated year with a population of 10,000 individuals within 5,000 households. NetLogo was originally designed for use in education so the programming language is very straightforward. Simple models can be set up within a few days of installing the software but the language is quite powerful and did not limit the functionality of the model. The package and language are well documented and there is an active user community for support if needed however I did not require any because the language was so easy to use. > > That seems to cover most of your requirements except that I have not implemented setting parameters from outside the program. I assume it could be done by reading in from a text file. Output to a stats package could be done by writing to a text file from the model them importing this into the external program. > > There was some talk a while ago about the authors moving towards making NetLogo open source but as far as I know, this has not been done yet. > > Hope this is helpful. The NetLogo website is http://ccl.northwestern.edu/netlogo/ <https://exchange5.essex.ac.uk/exchweb/bin/redir.asp?URL=http://ccl.northwestern.edu/netlogo/> . There is an early prototype of my model at http://istr.essex.ac.uk/tasc/students/tlawso/model2/ <https://exchange5.essex.ac.uk/exchweb/bin/redir.asp?URL=http://istr.essex.ac.uk/tasc/students/tlawso/model2/> > > Regards, > > Tony Lawson > > > ________________________________ > > From: News and discussion about computer simulation in the social sciences on behalf of Oliver Mannion > Sent: Thu 25/06/2009 22:06 > To: [log in to unmask] > Subject: [SIMSOC] Microsimulation Software Selection > > > > Hi there, > > We are a University based research centre embarking on the development of a > microsimulation model of health outcomes using empirical longitudinal data. > The end product will be a software program for use by external end users. > Our complete list of requirements is attached. > > We are currently in the software selection phase. Modgen and AnyLogic are > impressive packages in features and scope and would allow us to do most of > what we want with the least effort "out of the box". We are concerned > however that being closed source neither will allow 100% flexibility. > > Of the open source packages, Repast Simphony and Ascape have caught our > attention as mature tool kits, although oriented towards Agent Based > Modelling. We would therefore need to extend them to include input of > parameter and base data files. Of these Repast has the higher public profile > and more widespread usage. However Repast offers a lot of functionality we > don't need, at the cost of quite a bit of complexity which makes it more > difficult to understand the code and extend it. > > We would be grateful for any responses to the following questions: > > 1) Has anyone used Ascape for microsimulation? Is it widely used? Was it a > good fit for your project? > 2) What tool kits, if any, have you used for microsimulation? Or have you > written your microsimulation without the use of a 3rd party tool kit/package? > 3) If you have used Repast for microsimulation, what has been your > experience with it? > > Thanks and regards, > > Oliver Mannion > Programmer > COMPASS - Centre of Methods and Policy Application in the Social Sciences > www.compass.auckland.ac.nz > The University of Auckland, New Zealand > > Phone +(649) 373 7999 ext 89760 >