3

I would like to use Netlogo's Behavior Space to output a raster of patch values at the very end of a Behavor Space experiment. Is this possible? I don't want the output at the end of each replication (as you get from "Final Commands") as my rasters are around 1.2 MB with just zero values. I will use a counter if necessary, but I am wondering if anyone knows a built-in solution. Thanks in advance!

Luke C
  • 10,081
  • 1
  • 14
  • 21
  • Condition on `behavior-space-run-number`. Since you apparently only want to do this once, it would probably be easiest to just use the first run. – Alan May 10 '17 at 14:09
  • Unfortunately I'm after a variable that changes over the course of the experiment, so I need to do this on the final rep. – Luke C May 10 '17 at 14:34
  • Well, BS will tell you have many runs there will be, so you could still condition your output on the run number. Separately, it seems fairly likely that propagating a variable across runs like this is a flaw in experimental design. (However, perhaps you are doing this for some kind of adaptive search.) – Alan May 10 '17 at 15:13
  • Yep, that's an option for sure, I was just wondering if there's an automatic way. To your second point- I meant variable in the coding sense only; it records patch use to make a simple heat map and does not affect the experiment at all (or indeed, get used in any kind of analysis beyond visual). I think I would just use Behaviorsearch to do any heuristic searching. – Luke C May 10 '17 at 16:04
  • This need matches well with something that I expressed as a comment to [this answer](https://stackoverflow.com/a/66764034/12391423), noticing that it would be useful to have the exact opposite too: a command to be run by BS only at the very beginning of the experiment. Would these two (same but opposite) features be worth adding? – Matteo Aug 25 '21 at 10:28

0 Answers0