Profile Forecasters

Profile Forecasters compute a rolling average forecast over a profile pointer. The implementation is class J_ProfileForecaster.

A forecaster takes a profile pointer and computes the average value over a configurable forecast horizon. As the simulation progresses, the forecast window slides forward and the average is updated incrementally.

The constructor accepts the following arguments:

• String forecastName: An optional name for this forecaster. If null, it is automatically generated from the profile pointer name and forecast horizon.
• J_ProfilePointer profilePointer: The underlying profile to forecast over.
• double forecastTime_h: The length of the forecast horizon in hours.
• double currentTime_h: The current simulation time used for the initial forecast calculation.
• double timeStep_h: The simulation time step, used to determine the number of samples in the forecast window.

Key methods:

• initializeForecast(double currentTime_h): Computes the initial forecast as the average of all profile values from currentTime_h to currentTime_h + forecastTime_h.
• updateForecast(double t_h): Updates the forecast by removing the contribution of the value at t_h and adding the contribution of the value at t_h + forecastTime_h (sliding window).
• getForecast(): Returns the current forecasted average value.
• getForecastTime_h(): Returns the length of the forecast horizon.
• getName(): Returns the name of this forecaster.

The sliding window update in updateForecast is efficient: it only needs two profile lookups per timestep rather than recomputing the entire average.