I am trying to use dynamic Eigen matrices and vectors with boost::odeint. The reason is that I want to solve a problem defined by user input of arbitrary size. The problem can be captured in a state space equation in the following form: x' = A*x + B*u
(https://en.wikipedia.org/wiki/State-space_representation), which is basically the dxdt() function required for boost::odeint.
Following code is an example of what I am trying do:
#include <iostream>
#include <Eigen/Core>
#include <cstdlib>
#include <boost/numeric/odeint.hpp>
#include <boost/numeric/odeint/external/eigen/eigen_algebra.hpp>
typedef Eigen::VectorXd state_type;
state_type x;
Eigen::MatrixXd A;
void ODE_function (const state_type &x, state_type &dxdt, double)
{
dxdt = A * x;
}
void write_states(const state_type &x, const double t)
{
std::cout << t << "\t";
for (int i = 0; i < x.size(); i++)
{
std::cout << *(x.data()+i) << "\t";
}
std::cout << std::endl;
}
int main()
{
int nr_of_states;
std::cout << "How many (random) states would you like to simulate?: ";
std::cin >> nr_of_states;
std::cout << std::endl;
x = state_type(nr_of_states);
A = Eigen::MatrixXd(nr_of_states, nr_of_states);
srand(365);
for (int i = 0; i < A.size(); i++)
{
*(A.data()+i) = ((double)rand()/(double)RAND_MAX);
}
for (int i = 0; i < x.size(); i++)
{
*(x.data()+i) = i;
}
typedef runge_kutta_dopri5<state_type, double, state_type, double, vector_space_algebra> stepper;
integrate_adaptive(stepper(), ODE_function, x, 0.0, 25.0, 0.1, write_states);
std::cout <<std::endl << "final state vector: " << std::endl << x << std::endl;
return 0;
}
The user can input the amount of states in the system, accordingly the state matrix A and the initial state values are assigned arbitrary values for brevity of the example. Boost::odeint is then used to solve the state space equation. Compiling above code fails, returning the following error message:
cc1plus.exe: out of memory allocating 174479 bytes
Everything compiles fine when I exclude the following two blocks of code:
#include <boost/numeric/odeint.hpp>
#include <boost/numeric/odeint/external/eigen/eigen_algebra.hpp>
and
typedef runge_kutta_dopri5<state_type, double, state_type, double, vector_space_algebra> stepper;
integrate_adaptive(stepper(), ODE_function, x, 0.0, 100.0, 1.0, write_states);
If now only <boost/numeric/odeint.hpp>
is included again, the same error message is returned by the compiler. I have tried to use UFT8 and UFT16 encoding on the source file, but to no avail.
I will be most gratefull to people that can help me with this problem with either suggestions or solutions
state_type
. It is strange that it happens withEigen::matrixXd
. Can you check how much memory your build is consuming? What kind of compiler are you using? – Halt