The Tetris Game Charger (TGC) is the result of the group project of C2.Enexis of class 23/24 of Jheronymous Academy of Data Science (JADS).
This package is a toy package which leverages simulation using Salabim to test multiple algorithms for scheduling electric vehicles (EV) in a charging facility.
Next to the normal priority rules, e.g FIFO, SPT, EDD we also implemented OLP (OnLine Linear Programming) and RL (Reinforcement Learning). The latter is however not available via this package.
| TGC members | Role | Quote |
|---|---|---|
| Alex Teeuwen | Reinforcement learning engineer | I taught my computer to learn from its mistakes |
| Anne-Marie Van Nes | Organizer and criticaster | my rubber ducks are in a row, but keep floating away |
| Dominique Fürst | Statistical distributions Expert | In the world of data, I fit in like an outlier at a mean party. |
| Floris Padt | OLP - engineer | I love linear solutions. If only life wasn't so nonlinear. |
| Henk Koster | Presentor & business owner | Presenting a use case: where optimism meets data science reality. |
The idea of OLP originates from this paper: Guo et al. - 2017 - Optimal online adaptive electric vehicle charging.
This packages use by default the glptk solver. Other solvers like mosek, ipopt (MH27), cplex or gurobi can be configured but do require a trial license setup.
- create new environment (recommended but optional)
pip install -i https://test.pypi.org/simple/ tgc-jads-2324
The next two links show how the final TGC model was built using Queueing theory and Linear Programming with the pyomo package.
The jupyter notebooks contain explanation and pyton coding with results.
The last link is a downloadable excel file containing a lite OLP model which can be solved by the Excel Solver Add-in (needs to be activated)