This repository contains the experiments that explore sensing-driven clustering techniques applied in Aggregate Computing.
In result discussion there is a legend that explain how to interpret the graphical simulations.
A working version of Java, the supported version range is 11 to 17, and a working version of Python 3, including pip.
To run the Alchemist simulations devised from this paper, you could
use the pre-defined gradle task runAllGraphic
.
This task will launch each simulation defined by a yaml file in src/main/yaml
.
Hence, to run all the simulation you have to write
./gradlew runAllGraphic
Press P to start the simulation.
For further information about the GUI, see the graphical interface shortcuts.
In Windows, I suggest you to use WSL 2. Otherwise, you should write:
gradlew.bat runAllGraphic
The experiment is entirely reproducible. Regenerating all the data may take weeks on a well-equipped 2021 personal computer. The process is CPU-intensive and we do not recommend running it on devices where heat can be damaging to the battery.
In order to re-run all the experiments, launch:
./gradlew runAllBatch
data will be generated into the data
folder
If you are just willing to re-run the data analysis and generate all the charts in the paper (plus hundreds other), you can use the data we generated in the past:
pip install -r requirements.txt --user
./plot-all-clusters.sh