MPC¶

class do_mpc.controller.MPC(model)[source]¶

Model predictive controller.

For general information on model predictive control, please read our background article.

The MPC controller extends the do_mpc.optimizer.Optimizer base class (which is also used for the do_mpc.estimator.MHE estimator).

Use this class to configure and run the MPC controller based on a previously configured do_mpc.model.Model instance.

Configuration and setup:

Configuring and setting up the MPC controller involves the following steps:

Use set_param() to configure the MPC instance.
Set the objective of the control problem with set_objective() and set_rterm()
Set upper and lower bounds with bounds (optional).
Set further (non-linear) constraints with set_nl_cons() (optional).
Use the low-level API (get_p_template() and set_p_fun()) or high level API (set_uncertainty_values()) to create scenarios for robust MPC (optional).
Finally, call setup().

Warning

Before running the controller, make sure to supply a valid initial guess for all optimized variables (states, algebraic states and inputs). Simply set the initial values of x0, z0 and u0 and then call set_initial_guess().

To take full control over the initial guess, modify the values of opt_x_num.

During runtime call make_step() with the current state \(x\) to obtain the optimal control input \(u\).

Attributes

`MPC.bounds`	Query and set bounds of the optimization variables.
`MPC.opt_p_num`	Full MPC parameter vector.
`MPC.opt_x_num`	Full MPC solution and initial guess.
`MPC.scaling`	Query and set scaling of the optimization variables.
`MPC.t0`	Current time marker of the class.
`MPC.terminal_bounds`	Query and set the terminal bounds for the states.
`MPC.u0`	Initial input and current iterate.
`MPC.x0`	Initial state and current iterate.
`MPC.z0`	Initial algebraic state and current iterate.

Methods

`MPC.get_p_template`	Obtain output template for `set_p_fun()`.
`MPC.get_tvp_template`	Obtain output template for `set_tvp_fun()`.
`MPC.make_step`	Main method of the class during runtime.
`MPC.reset_history`	Reset the history of the optimizer.
`MPC.set_initial_guess`	Initial guess for optimization variables.
`MPC.set_nl_cons`	Introduce new constraint to the class.
`MPC.set_objective`	Sets the objective of the optimal control problem (OCP).
`MPC.set_p_fun`	Set function which returns parameters.
`MPC.set_param`	Set the parameters of the `MPC` class.
`MPC.set_rterm`	Set the penality factor for the inputs.
`MPC.set_tvp_fun`	Set function which returns time-varying parameters.
`MPC.set_uncertainty_values`	Define scenarios for the uncertain parameters.
`MPC.setup`	Setup the MPC class.
`MPC.solve`	Solves the optmization problem.

This page is auto-generated. Page source is not available on Github.