Class template explicit_stepper_base

boost::numeric::odeint::explicit_stepper_base — Base class for explicit steppers without step size control and without dense output.

Synopsis

// In header: <boost/numeric/odeint/stepper/base/explicit_stepper_base.hpp>

template<typename Stepper, unsigned short Order, typename State, 
         typename Value, typename Deriv, typename Time, typename Algebra, 
         typename Operations, typename Resizer> 
class explicit_stepper_base :
  public boost::numeric::odeint::algebra_stepper_base< Algebra, Operations >
{
public:
  // types
  typedef State                                       state_type;               
  typedef Value                                       value_type;               
  typedef Deriv                                       deriv_type;               
  typedef Time                                        time_type;                
  typedef Resizer                                     resizer_type;             
  typedef Stepper                                     stepper_type;             
  typedef stepper_tag                                 stepper_category;         
  typedef algebra_stepper_base< Algebra, Operations > algebra_stepper_base_type;
  typedef algebra_stepper_base_type::algebra_type     algebra_type;             
  typedef algebra_stepper_base_type::operations_type  operations_type;          
  typedef unsigned short                              order_type;               

  // public member functions
  template<typename System, typename StateInOut> 
    void do_step(System, StateInOut &, time_type, time_type);
  template<typename System, typename StateInOut> 
    void do_step(System, const StateInOut &, time_type, time_type);
  template<typename System, typename StateInOut, typename DerivIn> 
    boost::disable_if< boost::is_same< DerivIn, time_type >, void >::type 
    do_step(System, StateInOut &, const DerivIn &, time_type, time_type);
  template<typename System, typename StateIn, typename StateOut> 
    void do_step(System, const StateIn &, time_type, StateOut &, time_type);
  template<typename System, typename StateIn, typename DerivIn, 
           typename StateOut> 
    void do_step(System, const StateIn &, const DerivIn &, time_type, 
                 StateOut &, time_type);
  template<typename StateIn> void adjust_size(const StateIn &);
  const algebra_type & algebra() const;

  // private member functions
  stepper_type & stepper(void);
  const stepper_type & stepper(void) const;
  template<typename StateIn> bool resize_impl(const StateIn &);
  template<typename System, typename StateInOut> 
    void do_step_v1(System, StateInOut &, time_type, time_type);

  // public data members
  static const order_type order_value;
   __pad0__;
};

Description

This class serves as the base class for all explicit steppers with algebra and operations. Step size control and error estimation as well as dense output are not provided. explicit_stepper_base is used as the interface in a CRTP (currently recurring template pattern). In order to work correctly the parent class needs to have a method do_step_impl( system , in , dxdt_in , t , out , dt ). This is method is used by explicit_stepper_base. explicit_stepper_base derives from algebra_stepper_base. An example how this class can be used is

* template< class State , class Value , class Deriv , class Time , class Algebra , class Operations , class Resizer >
* class custom_euler : public explicit_stepper_base< 1 , State , Value , Deriv , Time , Algebra , Operations , Resizer >
* {
*  public:
*     
*     typedef explicit_stepper_base< 1 , State , Value , Deriv , Time , Algebra , Operations , Resizer > base_type;
*
*     custom_euler( const Algebra &algebra = Algebra() ) { }
* 
*     template< class Sys , class StateIn , class DerivIn , class StateOut >
*     void do_step_impl( Sys sys , const StateIn &in , const DerivIn &dxdt , Time t , StateOut &out , Time dt )
*     {
*         m_algebra.for_each3( out , in , dxdt , Operations::scale_sum2< Value , Time >( 1.0 , dt );
*     }
*
*     template< class State >
*     void adjust_size( const State &x )
*     {
*         base_type::adjust_size( x );
*     }
* };
*

For the Stepper concept only the do_step( sys , x , t , dt ) needs to be implemented. But this class provides additional do_step variants since the stepper is explicit. These methods can be used to increase the performance in some situation, for example if one needs to analyze dxdt during each step. In this case one can use

* sys( x , dxdt , t );
* stepper.do_step( sys , x , dxdt , t , dt );  // the value of dxdt is used here
* t += dt;
*

In detail explicit_stepper_base provides the following do_step variants

do_step( sys , x , t , dt ) - The classical do_step method needed to fulfill the Stepper concept. The state is updated in-place. A type modelling a Boost.Range can be used for x.
do_step( sys , in , t , out , dt ) - This method updates the state out-of-place, hence the result of the step is stored in out.
do_step( sys , x , dxdt , t , dt ) - This method updates the state in-place, but the derivative at the point t must be explicitly passed in dxdt. For an example see the code snippet above.
do_step( sys , in , dxdt , t , out , dt ) - This method update the state out-of-place and expects that the derivative at the point t is explicitly passed in dxdt. It is a combination of the two do_step methods above.

Note

The system is always passed as value, which might result in poor performance if it contains data. In this case it can be used with boost::ref or std::ref, for example stepper.do_step( boost::ref( sys ) , x , t , dt );

The time t is not advanced by the stepper. This has to done manually, or by the appropriate integrate routines or iterators.

Template Parameters

```
typename Stepper
```
The stepper on which this class should work. It is used via CRTP, hence explicit_stepper_base provides the interface for the Stepper.
```
unsigned short Order
```
The order of the stepper.
```
typename State
```
The state type for the stepper.
```
typename Value
```
The value type for the stepper. This should be a floating point type, like float, double, or a multiprecision type. It must not necessary be the value_type of the State. For example the State can be a vector< complex< double > > in this case the Value must be double. The default value is double.
```
typename Deriv
```
The type representing time derivatives of the state type. It is usually the same type as the state type, only if used with Boost.Units both types differ.
```
typename Time
```
The type representing the time. Usually the same type as the value type. When Boost.Units is used, this type has usually a unit.
```
typename Algebra
```
The algebra type which must fulfill the Algebra Concept.
```
typename Operations
```
The type for the operations which must fulfill the Operations Concept.
```
typename Resizer
```
The resizer policy class.

`explicit_stepper_base` public member functions

template<typename System, typename StateInOut> 
  void do_step(System system, StateInOut & x, time_type t, time_type dt);

This method performs one step. It transforms the result in-place.

Parameters:

`dt`	The step size.
`system`	The system function to solve, hence the r.h.s. of the ordinary differential equation. It must fulfill the Simple System concept.
`t`	The value of the time, at which the step should be performed.
`x`	The state of the ODE which should be solved. After calling do_step the result is updated in x.

template<typename System, typename StateInOut> 
  void do_step(System system, const StateInOut & x, time_type t, time_type dt);

Second version to solve the forwarding problem, can be called with Boost.Range as StateInOut.

template<typename System, typename StateInOut, typename DerivIn> 
  boost::disable_if< boost::is_same< DerivIn, time_type >, void >::type 
  do_step(System system, StateInOut & x, const DerivIn & dxdt, time_type t, 
          time_type dt);

The method performs one step. Additionally to the other method the derivative of x is also passed to this method. It is supposed to be used in the following way:

* sys( x , dxdt , t );
* stepper.do_step( sys , x , dxdt , t , dt );
*

The result is updated in place in x. This method is disabled if Time and Deriv are of the same type. In this case the method could not be distinguished from other do_step versions.

	Note
	This method does not solve the forwarding problem.

Parameters:

`dt`	The step size.
`dxdt`	The derivative of x at t.
`system`	The system function to solve, hence the r.h.s. of the ODE. It must fulfill the Simple System concept.
`t`	The value of the time, at which the step should be performed.
`x`	The state of the ODE which should be solved. After calling do_step the result is updated in x.

template<typename System, typename StateIn, typename StateOut> 
  void do_step(System system, const StateIn & in, time_type t, StateOut & out, 
               time_type dt);

The method performs one step. The state of the ODE is updated out-of-place.

	Note
	This method does not solve the forwarding problem.

Parameters:

`dt`	The step size.
`in`	The state of the ODE which should be solved. in is not modified in this method
`out`	The result of the step is written in out.
`system`	The system function to solve, hence the r.h.s. of the ODE. It must fulfill the Simple System concept.
`t`	The value of the time, at which the step should be performed.

template<typename System, typename StateIn, typename DerivIn, 
         typename StateOut> 
  void do_step(System system, const StateIn & in, const DerivIn & dxdt, 
               time_type t, StateOut & out, time_type dt);

The method performs one step. The state of the ODE is updated out-of-place. Furthermore, the derivative of x at t is passed to the stepper. It is supposed to be used in the following way:

* sys( in , dxdt , t );
* stepper.do_step( sys , in , dxdt , t , out , dt );
*

	Note
	This method does not solve the forwarding problem.

Parameters:

`dt`	The step size.
`dxdt`	The derivative of x at t.
`in`	The state of the ODE which should be solved. in is not modified in this method
`out`	The result of the step is written in out.
`system`	The system function to solve, hence the r.h.s. of the ODE. It must fulfill the Simple System concept.
`t`	The value of the time, at which the step should be performed.

```
template<typename StateIn> void adjust_size(const StateIn & x);
```
Adjust the size of all temporaries in the stepper manually.

Parameters:

x

A state from which the size of the temporaries to be resized is deduced.
```
const algebra_type & algebra() const;
```
Returns:

A reference to the algebra which is held by this class.

A const reference to the algebra which is held by this class.

`explicit_stepper_base` private member functions

```
stepper_type & stepper(void);
```

const stepper_type & stepper(void) const;

template<typename StateIn> bool resize_impl(const StateIn & x);

template<typename System, typename StateInOut> 
  void do_step_v1(System system, StateInOut & x, time_type t, time_type dt);

Boost C++ Libraries