Proof of Concept Replication - APGG-Lab/APGG GitHub Wiki
First Proof of Concept Experiment
To replicate the Proof of Concept Results first run APGG with this setup (Getting Started, How to run APGG):
"buildConfig.py" is by default configured for the first proof of concept experiment.
| Parameters | synergyFactor ($r$) | mutationRate ($\mu$) | Pop. Size | Group Size | eliminationCount | payoffIndividualism | punishmentCost ($\gamma$) | punishmentFine ($\beta$) | numGenerations | num_experiments | Mutator | Selector | Repopulator |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Value: | 1.5 to 7.0 (steps of 0.125) | 0.02 | 1000 (width: 100, height: 10) | 5 | 100 | 1.0 | 0.2 | 0.8 | 100,000 | 100 | Random Mutator | Random Selector | Proportionate Repopulator |
This will the all be saved in the setup.npy file and later used for plotting. This file can be ignored but not deleted or modified for successfull plotting!
Once the Experiment has finished running navigate into the experiments folder where all the data is stored! Within this folder you will find a set of Python Scripts and the Experiment data of the aforementioned experiment. Run the Python Scripts with said data and you should be able to see a graph similar to this one (Plotting Data):
Second Proof of Concept Experiment
To replicate the Proof of Concept Results first run APGG with this setup (Getting Started, How to run APGG):
| Parameters | synergyFactor ($r$) | mutationRate ($\mu$) | Pop. Size | Group Size | eliminationCount | payoffIndividualism | punishmentCost ($\gamma$) | punishmentFine ($\beta$) | numGenerations | num_experiments | Mutator | Selector | Repopulator |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Value: | 1.5 to 7.0 (steps of 0.125) | 0.02 | 1000 (width: 100, height: 10) | 5 | 100 | 0.5 | 0.0 | 0.0 | 100,000 | 100 | Random Mutator | Random Selector | Proportionate Repopulator |
This will the all be saved in the setup.npy file and later used for plotting. This file can be ignored but not deleted or modified for successfull plotting!
Once the Experiment has finished running navigate into the experiments folder where all the data is stored! Within this folder you will find a set of Python Scripts and the Experiment data of the aforementioned experiment. Run the Python Scripts with said data and you should be able to see a graph similar to this one (Plotting Data):
Additional Remarks
- "buildConfig.py" is by default configured for the first proof of concept experiment.
- Both proof of concept experiment may require plenty of computing time on your machine, it is therefore recommended to, if possible, run APGG on multiple cores. This can be done by specifying the number of cores when starting the code in the command line. For example "apgg.exe 12" would start APGG on 12 cores and "apgg.exe 4" would start it on 4, respectively.