One thought on “Linking Netlogo with R

  1. Dear Duncan,

    would be interesting to see the details from which the conclusion that data transfer rates with RNetLogo are slow was drawn.

    There are some many different things influencing the data transfer rates that I think such a general conclusion could be wrong.

    For example, it depends on the data structure: nested lists are complex and will take more time,
    data.frames and lists perform different.
    Furthermore, single value transformations are much slower than list transformations.
    And, very important, how was the JVM started! There are several options to give more power to the JVM (by specifying JVM startup arguments) as documented in the manual pages of NLStart.

    There are two vignettes to RNetLogo, which are of interested to this topic:
    Performance Notes: http://cran.r-project.org/web/packages/RNetLogo/vignettes/performanceNotes.pdf
    and
    Parallel Processing with RNetLogo: http://cran.r-project.org/web/packages/RNetLogo/vignettes/parallelProcessing.pdf

    See, for example, Fig. 5 and 6 in Perforamnce Notes:
    400,000 values in 0.35 sec.
    or
    1,200,000 values in 0.7 sec.
    I think, this is fast!

    And, these results were generated with a rather old notebook bought in year 2008 (Core(TM)2 Duo T9500, 32-bit Win XP, 4GB Ram) without specific JVM arguments (default values)!

    Best,
    Jan

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s