Skip to content

Latest commit

 

History

History
42 lines (32 loc) · 2.59 KB

README.md

File metadata and controls

42 lines (32 loc) · 2.59 KB
Raw

SimpleCVODEWrapper

A C library that wraps CVODE in a simpler interface for solving system of differential equations.

The following steps are necessary to succesfully accomplish the integration of a system of ordinary differential equations.

Define your system function

You should declare a function f that has the signature int f(realtype t, N_Vector y, N_Vector ydot, void *f_data).

You can get a definition of the type N_Vector by including the library nvector, #include<nvector/nvector_serial.h>.

Create you solver

Call the function new_cvode_solver(lmm) to start a SimpleCVODESolver object. The argument lmm stands for linear multistep method and can be either set as STIFF_INTEGRATOR (which uses BDF method), or NONSTIFF_INTEGRATOR (which uses Adams method).

Initialize the solver

Use the function init_solver(SimpleCVODESolver *solver, void *f, float t0, float *y0, int n), initialize the solver solver. The param f should point to your system function; t0 should be the starting time for integration, y0 is an array that determines the initial conditions; and n is the size of y0.

Set integration error tolerances parameters

Use the function set_tolerance(SimpleCVODESolver *solver, float abstol, float reltol) to define the integration error tolerance.

Prepare the solver for integration

The function prepare_solver creates data structures necessary for the integrator to work. These structures are related to the estimation of the Jacobian matrix, which is set to be approximated through difference quotients.

Define system optional arguments

It is possible to make the integrator pass arguments to the system function. To define these argments, the function set_system_data(SimpleCVODESolver *solver, void *data) should be called.

Integrate

Finally, one can integrate the system by calling the function integrate(SimpleCVODESolver *solver, float *t, int m). The times for which the integration values should be stored are defined in the array t, which should have size m. This function returns a matrix of size m x n where n is the cardinality of the system. On this matrix, each line has the integrated values of the system on a specific time step.

After integration

Once you have integrated your system, you can restart it by calling the function reset_solver. This function is useful for repeating the integration process by changing optional system parameters or even initial conditions. When you are done with integrations, use the function delete_solver to free all allocated memory related to the solver.