A depot charges its electric fleet over 24 hours. Each hour we choose an aggregate charging power, capped by the site's grid connection. Enough energy must arrive by noon (the morning-shift vehicles) and a larger total by end of day, while the electricity price swings from cheap overnight to an evening peak. The optimizer front-loads cheap overnight energy to clear the noon deadline and tops up the rest in the cheapest remaining hours. Score is total cost plus deadline penalties, to minimise. 24-D continuous problem.
| Algorithm | Cost+pen | Schedules | Deadlines |
|---|---|---|---|
| — no runs yet — | |||
Each hour the depot draws a charging power in [0, 10] (grid cap). The fleet needs
45 units of energy delivered by hour 12 and 110 units by hour 24. The objective is
the price-weighted energy cost plus a penalty of 50 per unit of unmet energy at
each deadline. The optimum charges hard in the cheap overnight hours to satisfy the
noon requirement, then fills the remaining need wherever the price is lowest,
backing off entirely during the evening peak. The JS objective on this page is a
line-for-line port of the Python example_applications/ev_fleet_charging
objective, so the two agree to floating-point tolerance.