All Adams solutions are time based, and are using implicit integration methods with adaptive timestep. The integration process is divided in a predictor and a corrector phase.
The predictor uses previous values for each state (and the current state derivative) to extrapolate to the next time. This prediction is entered into the equation system and an error can be estimated. If this error is too large, another order (number of points used for prediction) will be tried.
If the error is small enough, the corrector phase will start. This is basically a Newton-Raphson iteration to converge all equations to be satisfactory close to zero.
That is the basis, but with a thousand small details involved of course.
All Adams solutions are time based, and are using implicit integration methods with adaptive timestep. The integration process is divided in a predictor and a corrector phase.
The predictor uses previous values for each state (and the current state derivative) to extrapolate to the next time. This prediction is entered into the equation system and an error can be estimated. If this error is too large, another order (number of points used for prediction) will be tried.
If the error is small enough, the corrector phase will start. This is basically a Newton-Raphson iteration to converge all equations to be satisfactory close to zero.
That is the basis, but with a thousand small details involved of course.
