Experiments for the journal paper about the graph-evolution tool.
Pipeline
-
Create objectives and target values with create_objectives.py
-
Check that the objectives are valid and reachable with check_objectives.py
-
Quantify the difficultly of reaching those objectives with sample_objectives.py
-
Find the best model parameters for a subset of the objectives with paramsweep_jobs.py
-
This will generate scripts for running the parameter sweep using graph-evolution/main.py on the MSU HPCC
-
Analyze the results with paramsweep_analysis.py
-
In progress
- Actual main experiments for paper (iteratively add objectives to runs)