Is NetLogo a good platform for big models (>10,000s of patches, turtles)?
How can I speed up a model that runs very slowly?
Is NetLogo a good platform for big models (>10,000s of patches, turtles)?
How can I speed up a model that runs very slowly?
We just published an article on execution speed of NetLogo; it is available at: http://jasss.soc.surrey.ac.uk/20/1/3.html
The article's main points are (a) NetLogo is not necessarily slow for executing large scientific models, and in fact has speed advantages over some alternatives; and (b) NetLogo models do often execute very slowly at first but can almost always be sped up dramatically by using a few basic techniques. We provide a step-by-step strategy for measuring and improving execution speed. We also identify tools for efficient analysis (sensitivity analysis, parameter-fitting, etc.) of NetLogo models.
I just posted two updates to our article on NetLogo execution speed. The original article is at http://jasss.soc.surrey.ac.uk/20/1/3.html
One update provides very important information from the NetLogo development team on when in-radius and in-cone are and are not slow. The other corrects a conclusion about the cost of using links. The updates are at: http://www.railsback-grimm-abm-book.com/JASSS-models.html
Steve Railsback