Skip to content

Dispatch Mode

By default, EPM uses representative days to reduce computation time. Dispatch mode (fDispatchMode=1) runs the full hourly chronology instead, enabling commitment-level constraints that are not captured in a representative-day run.

Use dispatch mode when you need to model:

  • Startup costs and minimum up/down times
  • Ramping constraints over the full time horizon
  • Storage continuity across hours and days
  • Unit commitment binaries (on/off status per generator per hour)

For the full mathematical formulation of dispatch equations, see Dispatch Mode Equations.

How to activate

Set the following flags in your pSettings.csv:

Flag Value Description
fDispatchMode 1 Required. Activates dispatch mode and full hourly chronology
fRampConstraints 1 Enable ramp rate constraints
fApplyRampConstraint 1 Apply ramp constraints in equations
fApplyMinGenerationConstraint 1 Enforce minimum generation — required to activate unit commitment
fApplyStartupCost 1 Include startup costs in the objective function
fApplyMUDT 1 Enforce minimum up/down time windows

Only fDispatchMode is strictly required. The other flags activate specific features within dispatch mode and can be combined as needed.

Before running

  • Make sure input.gdx contains the full chronology: sets AT, pHours, and mapTS(q,d,t,AT)
  • If you changed any flag in pSettings.csv, re-run Python input treatment so the updated scalars are written to input.gdx
  • Test with a short time horizon first — dispatch mode is significantly more compute-intensive than a representative-day run

Run

Same command as a standard run — GAMS picks up fDispatchMode from pSettings.csv automatically:

python epm.py --folder_input your_data --config your_data/config.csv