Control¶

class mpcpy.Control(stateestimation, prediction, parameters=None, horizon=None, timestep=None, receding=None, savesolutions=0)[source]¶

Base class for defining the control for an mpc simulation

the formulation and solution methods must be redefined in a child class to generate control signals over the control horizon

Initializes the control object

Parameters:

stateestimation : mpcpy.Stateestimation

The object used to determine the state at the beginning of the control horizon.

prediction : mpcpy.Prediction object

The object used to determine the predictions over the control horizon.

parameters : dict

Dictionary specifying control parameters.

horizon : number

The length of the control horizon.

timestep : number

The length of the control timestep.

receding : number

The receding time, i.e. the time between subsequent calls to this control object.

savesolutions : int

Number of control solutions to be saved in the control object. Set to -1 to save all solutions.

__call__(starttime)[source]¶

Calculate the value of the control signal over the control horizon. Calls the solution method.

Parameters:

starttime : real

Time at the beginning of the control horizon.

Returns:

dict

A dictionary with representing the control signals with time.

formulation()[source]¶

Performs actions the first time the control is run.

Can be used to set up an optimal control problem. The set-up and solution are separated as sometimes the set-up takes a significant amount of time and performs actions which must not be repeated.

Should be redefined in a child class to set up the actual control algorithm.

solution(state, prediction)[source]¶

Returns the control profiles (“the plan”).

Should be redefined in a child class to set up the actual control algorithm.

Parameters:

state : dict

Dictionary with the states at the start of the control horizon as returned by the stateestimation object.

prediction : dict

Dictionary with the values of predictions over the control horizon as returned by the prediction object.

Returns:

dict

Dictionary with the solution, “time” must be a key.

time(starttime)[source]¶

Returns a time vector over the control horizon with the defined timestep

Parameters:

starttime : real

Time at the beginning of the control horizon.