Stan Math Library  2.10.0
reverse mode automatic differentiation
integrate_ode_rk45.hpp
Go to the documentation of this file.
1 #ifndef STAN_MATH_PRIM_ARR_FUNCTOR_INTEGRATE_ODE_RK45_HPP
2 #define STAN_MATH_PRIM_ARR_FUNCTOR_INTEGRATE_ODE_RK45_HPP
3 
13 #include <boost/numeric/odeint.hpp>
14 #include <ostream>
15 #include <vector>
16 
17 namespace stan {
18 
19  namespace math {
20 
65  template <typename F, typename T1, typename T2>
66  std::vector<std::vector<typename stan::return_type<T1, T2>::type> >
67  integrate_ode_rk45(const F& f,
68  const std::vector<T1> y0,
69  const double t0,
70  const std::vector<double>& ts,
71  const std::vector<T2>& theta,
72  const std::vector<double>& x,
73  const std::vector<int>& x_int,
74  std::ostream* msgs = 0,
75  double relative_tolerance = 1e-6,
76  double absolute_tolerance = 1e-6,
77  int max_num_steps = 1E6) {
78  using boost::numeric::odeint::integrate_times;
79  using boost::numeric::odeint::make_dense_output;
80  using boost::numeric::odeint::runge_kutta_dopri5;
81  using boost::numeric::odeint::max_step_checker;
82 
83  check_finite("integrate_ode_rk45", "initial state", y0);
84  check_finite("integrate_ode_rk45", "initial time", t0);
85  check_finite("integrate_ode_rk45", "times", ts);
86  check_finite("integrate_ode_rk45", "parameter vector", theta);
87  check_finite("integrate_ode_rk45", "continuous data", x);
88 
89  check_nonzero_size("integrate_ode_rk45", "times", ts);
90  check_nonzero_size("integrate_ode_rk45", "initial state", y0);
91  check_ordered("integrate_ode_rk45", "times", ts);
92  check_less("integrate_ode_rk45", "initial time", t0, ts[0]);
93 
94  if (relative_tolerance <= 0)
95  invalid_argument("integrate_ode_rk45",
96  "relative_tolerance,", relative_tolerance,
97  "", ", must be greater than 0");
98  if (absolute_tolerance <= 0)
99  invalid_argument("integrate_ode_rk45",
100  "absolute_tolerance,", absolute_tolerance,
101  "", ", must be greater than 0");
102  if (max_num_steps <= 0)
103  invalid_argument("integrate_ode_rk45",
104  "max_num_steps,", max_num_steps,
105  "", ", must be greater than 0");
106 
107  // creates basic or coupled system by template specializations
109  coupled_system(f, y0, theta, x, x_int, msgs);
110 
111  // first time in the vector must be time of initial state
112  std::vector<double> ts_vec(ts.size() + 1);
113  ts_vec[0] = t0;
114  for (size_t n = 0; n < ts.size(); n++)
115  ts_vec[n+1] = ts[n];
116 
117  std::vector<std::vector<double> > y_coupled(ts_vec.size());
118  coupled_ode_observer observer(y_coupled);
119 
120  // the coupled system creates the coupled initial state
121  std::vector<double> initial_coupled_state
122  = coupled_system.initial_state();
123 
124  const double step_size = 0.1;
125  integrate_times(make_dense_output(absolute_tolerance,
126  relative_tolerance,
127  runge_kutta_dopri5<std::vector<double>,
128  double,
129  std::vector<double>,
130  double>() ),
131  coupled_system,
132  initial_coupled_state,
133  boost::begin(ts_vec), boost::end(ts_vec),
134  step_size,
135  observer,
136  max_step_checker(max_num_steps));
137 
138  // remove the first state corresponding to the initial value
139  y_coupled.erase(y_coupled.begin());
140 
141  // the coupled system also encapsulates the decoupling operation
142  return coupled_system.decouple_states(y_coupled);
143  }
144 
145  }
146 
147 }
148 
149 #endif
bool check_less(const char *function, const char *name, const T_y &y, const T_high &high)
Return true if y is strictly less than high.
Definition: check_less.hpp:81
bool check_nonzero_size(const char *function, const char *name, const T_y &y)
Return true if the specified matrix/vector is of non-zero size.
bool check_ordered(const char *function, const char *name, const std::vector< T_y > &y)
Return true if the specified vector is sorted into strictly increasing order.
Observer for the coupled states.
std::vector< std::vector< typename stan::return_type< T1, T2 >::type > > integrate_ode_rk45(const F &f, const std::vector< T1 > y0, const double t0, const std::vector< double > &ts, const std::vector< T2 > &theta, const std::vector< double > &x, const std::vector< int > &x_int, std::ostream *msgs=0, double relative_tolerance=1e-6, double absolute_tolerance=1e-6, int max_num_steps=1E6)
Return the solutions for the specified system of ordinary differential equations given the specified ...
void invalid_argument(const char *function, const char *name, const T &y, const char *msg1, const char *msg2)
Throw an invalid_argument exception with a consistently formatted message.
double e()
Return the base of the natural logarithm.
Definition: constants.hpp:95
bool check_finite(const char *function, const char *name, const T_y &y)
Return true if y is finite.
Base template class for a coupled ordinary differential equation system, which adds sensitivities to ...

     [ Stan Home Page ] © 2011–2016, Stan Development Team.