agentlib_mpc.optimization_backends.casadi_ package

Subpackages

• agentlib_mpc.optimization_backends.casadi_.core package
  • Submodules
  • agentlib_mpc.optimization_backends.casadi_.core.VariableGroup module
    • OptimizationParameter
      • OptimizationParameter.add_default_values
      • OptimizationParameter.declare()
      • OptimizationParameter.dim
      • OptimizationParameter.full_names
      • OptimizationParameter.full_symbolic
      • OptimizationParameter.full_with_defaults
      • OptimizationParameter.name
      • OptimizationParameter.ref_names
      • OptimizationParameter.use_in_stage_function
    • OptimizationQuantity
      • OptimizationQuantity.dim
      • OptimizationQuantity.full_names
      • OptimizationQuantity.full_symbolic
      • OptimizationQuantity.name
      • OptimizationQuantity.ref_names
      • OptimizationQuantity.use_in_stage_function
    • OptimizationVariable
      • OptimizationVariable.binary
      • OptimizationVariable.declare()
      • OptimizationVariable.dim
      • OptimizationVariable.full_names
      • OptimizationVariable.full_symbolic
      • OptimizationVariable.input_map
      • OptimizationVariable.name
      • OptimizationVariable.output_map
      • OptimizationVariable.ref_names
      • OptimizationVariable.use_in_stage_function
  • agentlib_mpc.optimization_backends.casadi_.core.casadi_backend module
    • CasADiBackend
      • CasADiBackend.config_type
      • CasADiBackend.cost_function
      • CasADiBackend.discretization
      • CasADiBackend.discretization_types
      • CasADiBackend.model
      • CasADiBackend.reset_setup_attributes()
      • CasADiBackend.save_result_df()
      • CasADiBackend.setup_optimization()
      • CasADiBackend.solve()
      • CasADiBackend.system
      • CasADiBackend.system_type
      • CasADiBackend.var_ref
    • CasadiBackendConfig
      • CasadiBackendConfig.build_batch_bat
      • CasadiBackendConfig.discretization_options
      • CasadiBackendConfig.do_jit
      • CasadiBackendConfig.model
      • CasadiBackendConfig.name
      • CasadiBackendConfig.overwrite_result_file
      • CasadiBackendConfig.results_file
      • CasadiBackendConfig.save_results
      • CasadiBackendConfig.solver
      • CasadiBackendConfig.validate_compile
  • agentlib_mpc.optimization_backends.casadi_.core.discretization module
    • Discretization
      • Discretization.add_constraint()
      • Discretization.add_opt_par()
      • Discretization.add_opt_var()
      • Discretization.binary_vars
      • Discretization.constraints
      • Discretization.constraints_lb
      • Discretization.constraints_ub
      • Discretization.create_nlp_in_out_mapping()
      • Discretization.equalities
      • Discretization.grid()
      • Discretization.initial_guess
      • Discretization.initialize()
      • Discretization.k
      • Discretization.mpc_opt_pars
      • Discretization.mpc_opt_vars
      • Discretization.nlp
      • Discretization.objective_function
      • Discretization.only_positive_times_in_results
      • Discretization.opt_pars
      • Discretization.opt_vars
      • Discretization.opt_vars_lb
      • Discretization.opt_vars_ub
      • Discretization.pred_time
      • Discretization.solve()
    • Results
      • Results.columns
      • Results.df
      • Results.grid
      • Results.matrix
      • Results.stats
      • Results.stats_line()
      • Results.variable_grid_indices
      • Results.variable_lookup()
      • Results.write_columns()
      • Results.write_stats_columns()
  • agentlib_mpc.optimization_backends.casadi_.core.system module
    • System
      • System.initialize()
      • System.parameters
      • System.quantities
      • System.variables

Submodules

agentlib_mpc.optimization_backends.casadi_.admm module

class agentlib_mpc.optimization_backends.casadi_.admm.ADMMCollocation(options: CasadiDiscretizationOptions)

Bases: DirectCollocation

class agentlib_mpc.optimization_backends.casadi_.admm.ADMMMultipleShooting(options: CasadiDiscretizationOptions)

Bases: MultipleShooting

class agentlib_mpc.optimization_backends.casadi_.admm.CasADiADMMBackend(config: dict)

Bases: CasADiBaseBackend, ADMMBackend

Class doing optimization of ADMM subproblems with CasADi.

cost_function: Callable | None

property coupling_grid

Returns the grid on which the coupling variables are discretized.

discretization: DiscretizationT

discretization_types: dict[DiscretizationMethod, Type[DiscretizationT]] = {DiscretizationMethod.collocation: <class 'agentlib_mpc.optimization_backends.casadi_.admm.ADMMCollocation'>, DiscretizationMethod.multiple_shooting: <class 'agentlib_mpc.optimization_backends.casadi_.admm.ADMMMultipleShooting'>}

it: int

model: ModelT

now: float

result_stats: list[str]

results: list[pd.DataFrame]

save_result_df(results: Results, now: float = 0)

Save the results of solve into a dataframe at each time step.

Example results dataframe:

value_type variable … lower variable T_0 T_0_slack … T_0_slack mDot_0 time_step … 2 0.000000 298.160000 NaN … NaN NaN

101.431499 297.540944 -149.465942 … -inf 0.0 450.000000 295.779780 -147.704779 … -inf 0.0 798.568501 294.720770 -146.645769 … -inf 0.0

Parameters:

• results –

• now –

Returns:

system: CasadiADMMSystem

system_type

alias of CasadiADMMSystem

var_ref: mpc_datamodels.VariableReference | None

class agentlib_mpc.optimization_backends.casadi_.admm.CasadiADMMSystem

Bases: FullSystem

algebraics: OptimizationVariable

controls: OptimizationVariable

cost_function: ca.MX

exchange_diff: OptimizationParameter

exchange_multipliers: OptimizationParameter

global_couplings: OptimizationParameter

initial_state: OptimizationParameter

initialize(model: CasadiModel, var_ref: VariableReference)

last_control: OptimizationParameter

local_couplings: OptimizationVariable

local_exchange: OptimizationVariable

model_constraints: Constraint

model_parameters: OptimizationParameter

multipliers: OptimizationParameter

non_controlled_inputs: OptimizationParameter

ode: ca.MX

outputs: OptimizationVariable

penalty_factor: OptimizationParameter

r_del_u: OptimizationParameter

states: OptimizationVariable

agentlib_mpc.optimization_backends.casadi_.basic module

class agentlib_mpc.optimization_backends.casadi_.basic.BaseSystem

Bases: System

algebraics: OptimizationVariable

controls: OptimizationVariable

cost_function: MX

initial_state: OptimizationParameter

initialize(model: CasadiModel, var_ref: VariableReference)

model_constraints: Constraint

model_parameters: OptimizationParameter

non_controlled_inputs: OptimizationParameter

ode: MX

outputs: OptimizationVariable

states: OptimizationVariable

class agentlib_mpc.optimization_backends.casadi_.basic.CasADiBaseBackend(config: dict)

Bases: CasADiBackend

Class doing optimization of ADMM subproblems with CasADi.

cost_function: Callable | None

discretization: DiscretizationT

discretization_types: dict[DiscretizationMethod, Type[DiscretizationT]] = {DiscretizationMethod.collocation: <class 'agentlib_mpc.optimization_backends.casadi_.basic.DirectCollocation'>, DiscretizationMethod.multiple_shooting: <class 'agentlib_mpc.optimization_backends.casadi_.basic.MultipleShooting'>}

model: ModelT

system: BaseSystem

system_type

alias of BaseSystem

var_ref: mpc_datamodels.VariableReference | None

class agentlib_mpc.optimization_backends.casadi_.basic.CollocationMatrices(order: int, root: numpy.ndarray, B: numpy.ndarray, C: numpy.ndarray, D: numpy.ndarray)

Bases: object

B: ndarray

C: ndarray

D: ndarray

order: int

root: ndarray

class agentlib_mpc.optimization_backends.casadi_.basic.DirectCollocation(options: CasadiDiscretizationOptions)

Bases: Discretization

initialize(system: BaseSystem, solver_factory: SolverFactory)

Initializes the trajectory optimization problem, creating all symbolic variables of the OCP, the mapping function and the numerical solver.

class agentlib_mpc.optimization_backends.casadi_.basic.MultipleShooting(options: CasadiDiscretizationOptions)

Bases: Discretization

initialize(system: BaseSystem, solver_factory: SolverFactory)

Initializes the trajectory optimization problem, creating all symbolic variables of the OCP, the mapping function and the numerical solver.

agentlib_mpc.optimization_backends.casadi_.casadi_admm_ml module

class agentlib_mpc.optimization_backends.casadi_.casadi_admm_ml.CasADiADMMBackend_NN(config: dict)

Bases: CasADiADMMBackend, CasADiBBBackend

Class doing optimization with an MLModel.

discretization_types: dict[DiscretizationMethod, Type[DiscretizationT]] = {DiscretizationMethod.multiple_shooting: <class 'agentlib_mpc.optimization_backends.casadi_.casadi_admm_ml.MultipleShootingADMMNN'>}

system: CasadiADMMNNSystem

system_type

alias of CasadiADMMNNSystem

class agentlib_mpc.optimization_backends.casadi_.casadi_admm_ml.CasadiADMMNNSystem

Bases: CasadiADMMSystem, CasadiMLSystem

In this class, the lags are determined by the trainer alone and the lags are saved in the serialized MLModel so that it doesn’t have to be defined in the model again

initialize(model: CasadiMLModel, var_ref: VariableReference)

property parameters: list[agentlib_mpc.optimization_backends.casadi_.core.VariableGroup.OptimizationParameter]

past_couplings: OptimizationParameter

past_exchange: OptimizationParameter

property quantities: list[Union[agentlib_mpc.optimization_backends.casadi_.core.VariableGroup.OptimizationParameter, agentlib_mpc.optimization_backends.casadi_.core.VariableGroup.OptimizationVariable]]

property sim_step_quantities: dict[str, Union[agentlib_mpc.optimization_backends.casadi_.core.VariableGroup.OptimizationParameter, agentlib_mpc.optimization_backends.casadi_.core.VariableGroup.OptimizationVariable]]

property variables: list[agentlib_mpc.optimization_backends.casadi_.core.VariableGroup.OptimizationVariable]

class agentlib_mpc.optimization_backends.casadi_.casadi_admm_ml.MultipleShootingADMMNN(options: CasadiDiscretizationOptions)

Bases: ADMMMultipleShooting, MultipleShooting_ML

max_lag: int

agentlib_mpc.optimization_backends.casadi_.casadi_ml module

class agentlib_mpc.optimization_backends.casadi_.casadi_ml.CasADiBBBackend(config: dict)

Bases: CasADiBaseBackend

Class doing optimization with a MLModel.

discretization_types: dict[DiscretizationMethod, Type[DiscretizationT]] = {DiscretizationMethod.multiple_shooting: <class 'agentlib_mpc.optimization_backends.casadi_.casadi_ml.MultipleShooting_ML'>}

get_lags_per_variable() → dict[str, float]

Returns the name of variables which include lags and their lag. The MPC module can use this information to save relevant past data of lagged variables

lag_collection: Dict[str, deque] = {}

max_lag: int

system: CasadiMLSystem

system_type

alias of CasadiMLSystem

class agentlib_mpc.optimization_backends.casadi_.casadi_ml.CasadiMLSystem

Bases: FullSystem

all_system_quantities() → dict[str, agentlib_mpc.optimization_backends.casadi_.core.VariableGroup.OptimizationQuantity]

initialize(model: CasadiMLModel, var_ref: FullVariableReference)

lags_dict: dict[str, int]

property max_lag: int

model: CasadiMLModel

sim_step: Function

class agentlib_mpc.optimization_backends.casadi_.casadi_ml.MultipleShooting_ML(options: CasadiDiscretizationOptions)

Bases: MultipleShooting

initialize(system: CasadiMLSystem, solver_factory: SolverFactory)

Initializes the trajectory optimization problem, creating all symbolic variables of the OCP, the mapping function and the numerical solver.

max_lag: int

agentlib_mpc.optimization_backends.casadi_.full module

class agentlib_mpc.optimization_backends.casadi_.full.CasADiFullBackend(config: dict)

Bases: CasADiBackend

Class doing optimization of ADMM subproblems with CasADi.

cost_function: Callable | None

discretization: DiscretizationT

discretization_types: dict[DiscretizationMethod, Type[DiscretizationT]] = {DiscretizationMethod.collocation: <class 'agentlib_mpc.optimization_backends.casadi_.full.DirectCollocation'>, DiscretizationMethod.multiple_shooting: <class 'agentlib_mpc.optimization_backends.casadi_.full.MultipleShooting'>}

model: ModelT

system: FullSystem

system_type

alias of FullSystem

var_ref: mpc_datamodels.VariableReference | None

class agentlib_mpc.optimization_backends.casadi_.full.DirectCollocation(options: CasadiDiscretizationOptions)

Bases: DirectCollocation

class agentlib_mpc.optimization_backends.casadi_.full.FullSystem

Bases: BaseSystem

algebraics: OptimizationVariable

controls: OptimizationVariable

cost_function: ca.MX

initial_state: OptimizationParameter

initialize(model: CasadiModel, var_ref: FullVariableReference)

last_control: OptimizationParameter

model_constraints: Constraint

model_parameters: OptimizationParameter

non_controlled_inputs: OptimizationParameter

ode: ca.MX

outputs: OptimizationVariable

r_del_u: OptimizationParameter

states: OptimizationVariable

class agentlib_mpc.optimization_backends.casadi_.full.MultipleShooting(options: CasadiDiscretizationOptions)

Bases: MultipleShooting

agentlib_mpc.optimization_backends.casadi_.mhe module

class agentlib_mpc.optimization_backends.casadi_.mhe.CollocationMatrices(order: int, root: numpy.ndarray, B: numpy.ndarray, C: numpy.ndarray, D: numpy.ndarray)

Bases: object

B: ndarray

C: ndarray

D: ndarray

order: int

root: ndarray

class agentlib_mpc.optimization_backends.casadi_.mhe.DirectCollocation(options: CasadiDiscretizationOptions)

Bases: Discretization

initialize(system: MHESystem, solver_factory: SolverFactory)

Initializes the trajectory optimization problem, creating all symbolic variables of the OCP, the mapping function and the numerical solver.

only_positive_times_in_results: bool = False

class agentlib_mpc.optimization_backends.casadi_.mhe.MHEBackend(config: dict)

Bases: CasADiBackend

Class doing optimization of ADMM subproblems with CasADi.

discretization_types: dict[DiscretizationMethod, Type[DiscretizationT]] = {DiscretizationMethod.collocation: <class 'agentlib_mpc.optimization_backends.casadi_.mhe.DirectCollocation'>}

static sample(trajectory: float | int | Series | list[Union[float, int]], grid: list | ndarray, current: float = 0, method: str = 'linear') → list

Obtain the specified portion of the trajectory.

Parameters:

• trajectory – The trajectory to be sampled. Scalars will be expanded onto the grid. Lists need to exactly match the provided grid. Otherwise, a list of tuples is accepted with the form ( timestamp, value). A dict with the keys ‘grid’ and ‘value’ is also accepted.

• current – start time of requested trajectory

• grid – target interpolation grid in seconds in relative terms (i.e. starting from 0 usually)

• method – interpolation method, currently accepted: ‘linear’, ‘spline’, ‘previous’

Returns:

Sampled list of values.

Takes a slice of the trajectory from the current time step with the specified length and interpolates it to match the requested sampling. If the requested horizon is longer than the available data, the last available value will be used for the remainder.

Raises:

• ValueError –

• TypeError –

system: MHESystem

system_type

alias of MHESystem

class agentlib_mpc.optimization_backends.casadi_.mhe.MHESystem

Bases: System

algebraics: OptimizationVariable

cost_function: MX

estimated_inputs: OptimizationVariable

estimated_parameters: OptimizationVariable

estimated_states: OptimizationVariable

initialize(model: CasadiModel, var_ref: MHEVariableReference)

known_inputs: OptimizationParameter

known_parameters: OptimizationParameter

measured_states: OptimizationParameter

model_constraints: Constraint

ode: MX

outputs: OptimizationVariable

agentlib_mpc.optimization_backends.casadi_.minlp module

class agentlib_mpc.optimization_backends.casadi_.minlp.CasADiMINLPBackend(config: dict)

Bases: CasADiBackend

Class doing optimization of ADMM subproblems with CasADi.

cost_function: Callable | None

discretization: DiscretizationT

discretization_types: dict[DiscretizationMethod, Type[DiscretizationT]] = {DiscretizationMethod.collocation: <class 'agentlib_mpc.optimization_backends.casadi_.minlp.DirectCollocation'>}

model: ModelT

system: CasadiMINLPSystem

system_type

alias of CasadiMINLPSystem

var_ref: mpc_datamodels.VariableReference | None

class agentlib_mpc.optimization_backends.casadi_.minlp.CasadiMINLPSystem

Bases: BaseSystem

algebraics: OptimizationVariable

binary_controls: OptimizationVariable

controls: OptimizationVariable

cost_function: ca.MX

initial_state: OptimizationParameter

initialize(model: CasadiModel, var_ref: MINLPVariableReference)

model_constraints: Constraint

model_parameters: OptimizationParameter

non_controlled_inputs: OptimizationParameter

ode: ca.MX

outputs: OptimizationVariable

states: OptimizationVariable

class agentlib_mpc.optimization_backends.casadi_.minlp.DirectCollocation(options: CasadiDiscretizationOptions)

Bases: DirectCollocation

agentlib_mpc.optimization_backends.casadi_.minlp_cia module