This repository proposes a structure to facilitate the use of the rDaNCES package to run agent-based simulations based on the DaNCES framework. Specifically, we recommend setting up:
-
A 'code' folder to include the scripts that run simulations and analyze the simulated data. See get_started.R script for a step by step guide.
-
The rDaNCES folder (currently included as a submodule).
-
A 'simulator' folder that includes the model executables (preferably the full bin folder) to be used (see example of sim_demo). For a trial in Windows, the user can insert the compiled Release found in the DaNCES_framework repo. To keep track in the README of each full model, the user is advised to take advantage of version control of their model's GitHub repository, for instance:
- demo_model.exe: branch develop @583ebdf (tree: 732qsfda4n, commit am: 'exp 1 model')
- demo_model_v2.exe: branch develop @535edde (tree: 732qsfda4n, commit am: 'exp 2 model')
The model will create a sim_data folder in the repository where it will store all simulated data in a subdirectory named as in the default config (here 'test') or set through the experiment set up (here in get_started.R set as 'rdances_demo').
Apart from the R dependencies included in the package's documentation, to run the agent-based simulations:
- See prerequesites of C++ model DaNCES
In Windows, using a precompiled model, you may also need to:
- Define an enviornmental variable for the path to the full dances repo
Further information on the functionality of the framework see our accompanying publication:
Papadopoulou M, Hildenbrandt H, Hemelrijk CK. (2025) A data-inspired framework to simulate collective behaviour of predator-prey systems. Adaptive Behavior, doi: 10.1177/10597123251364750.
The development of this software was mainly supported by an NWO grant awarded to Prof. Dr. Charlotte K. Hemelrijk at the University of Groningen.
- Marina Papadopoulou - For any problem email at: m.papadopoulou.rug@gmail.com