We are still working on the code, but we provide a working codebase as it is. We are working on a github repo with much more clean code and optimized experimental structure.
.
├── logs
│ ├── read_acc.py
│ └── ...
├── README.md
├── requirements.txt
└── src
├── optimal_exp
│ ├── ... <--- all scripts for figure 2 and all tables (you need to run manually each one)
├── res_consumption_exp
│ ├── ... <--- all scripts for figure 3 (run.sh to run all)
├── supplementary
├── ... <--- all scripts for supplementary (you need to run manually each one)
We suggest the following steps (for linux):
-
open a terminal cd final_code
-
install virtualenv then:
a.
virtualenv venv
b.source venv/bin/activate
c.pip install -r requirements
-
To reproduce figure3
cd src/res_consumption_exp/fig3/
a.
run.sh && check.sh
it will take days -
To reproduce figure2
cd src/optimal_exp/
a. you can cd into whatever folder and run an experimental point of figure2 and tables, the names of the folder are self explanatory.