Class ArkExplicitIntegrator

Class ArkExplicitIntegrator#

Nested Relationships#

Nested Types#

Inheritance Relationships#

Base Type#

Class Documentation#

class ArkExplicitIntegrator : public Karana::Math::Integrator#

The multistep ArkODE explicit integrator from the SUNDIALS suite.

Public Types

enum class NLSOLVER#

the nonlinear solver type to use

Values:

enumerator FIXEDPOINT#
enumerator NEWTON#
enum class METHOD#

the solver type to use

Values:

enumerator HUEN_EULER#
enumerator ZONNEVELD#
enumerator CASH_KARP#
enumerator DORMAND_PRINCE#
enumerator FEHLBERG78#

Public Functions

ArkExplicitIntegrator(size_t nstates, DerivativeFunction f, const IntegratorType &integrator_type, const IntegratorOptions *options = nullptr)#

Constructor.

Parameters:

  • nstates – The number of states to integrate.

  • f – The derivative function to calculate the derivative of the states.

  • integrator_type – The type of ark integrator to use.

  • options – The integrator options.

~ArkExplicitIntegrator()#
virtual std::string_view typeString() const override#

Return the integrator type as a string.

Returns:

the integrator type

virtual std::unique_ptr<Integrator::IntegratorOptions> getOptions() override#

Get the options associated with the integrator.

Returns:

The options associated with the integrator.

virtual IntegratorType getIntegratorType() const override#

Get the IntegratorType for this integrator.

Returns:

The IntegratorType for this integrator.

virtual void hardReset(size_t nstates) override#

Reset the state size.

Parameters:

nstates – the new state size

virtual void softReset(const Ktime &new_t, const Vec &new_x) override#

Reset the cached time and state and to the new values going forward.

Parameters:

  • new_t – the new time

  • new_x – the new state

virtual Ktime advanceTo(const Ktime &t_end) override#

Method to advance the sytem state to a specified time.

Parameters:

to_t – the desired end time

Returns:

the actual end time

Public Members

const std::map<METHOD, ARKODE_ERKTableID> butcher_table_id{{METHOD::HUEN_EULER, ARKODE_HEUN_EULER_2_1_2}, {METHOD::ZONNEVELD, ARKODE_ZONNEVELD_5_3_4}, {METHOD::CASH_KARP, ARKODE_CASH_KARP_6_4_5}, {METHOD::DORMAND_PRINCE, ARKODE_DORMAND_PRINCE_7_4_5}, {METHOD::FEHLBERG78, ARKODE_FEHLBERG_13_7_8}}#

map with the Butcher table number (from arkode_butcher_erk.h) for the different integrator choices. The numbers are the integrator stages/embedding order/order

Protected Attributes

SUNContext _sunctx = nullptr#

The Sundials context object.

N_Vector _y = nullptr#

The current integrator state.

SUNMatrix _a_mat = nullptr#

Matrix used in the linear solver.

SUNLinearSolver _ls = nullptr#

Linear solver.

SUNNonlinearSolver _nls = nullptr#

Nonlinear solver.

void *_arkode_mem#

The Arkode memory.

Vec _dx#

The current derivative of the state.

Protected Static Functions

static int _arkode_f(realtype t, N_Vector y, N_Vector ydot, void *user_data)#

Calles the derivative function on the provided integrator state, and puts the derivative of the state into ydot.

Parameters:

  • t – The time to use when calculating the derivative.

  • y – The current integrator state.

  • ydot – The vector that will get populated with the derivative of the state.

  • user_data – Extra data used by the deriative function.

Returns:

An error code indicating if the call was successful or not.

struct IntegratorOptions : public Karana::Math::Integrator::IntegratorOptions#

Struct with options for this integrator

Public Functions

IntegratorOptions &operator=(const IntegratorOptions &p)#

Assignment operator.

Parameters:

p – The IntegratorOptions to copy values from.

Returns:

A reference to the newly assigned IntegratorOptions.

IntegratorOptions(const IntegratorOptions &p)#

Copy constructor.

Parameters:

p – The IntegratorOptions to copy values from.

IntegratorOptions() = default#

Constructor.

virtual ~IntegratorOptions()#

Public Members

double rtol = 1e-4#

The desired relative tolerance used to control the integrator error.

double atol = 1e-8#

The desired absolute tolerance used to control the integrator error.

Vec atol_vec#

The desired absolute tolerance vector used to control the integrator error. This can be used rather than a scalar tolerance to set different values for different states.

long max_num_steps = 500#

Maximum number of steps the solver can take before erroring.

NLSOLVER nlsolver = NLSOLVER::FIXEDPOINT#

The Nonlinear solver type to use.

METHOD method = METHOD::ZONNEVELD#

The solver method to use.

Contents