Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Relative tolerance vs Relative Error
Posted Mar 17, 2016, 11:19 p.m. EDT Parameters, Variables, & Functions, Studies & Solvers 2 Replies
Please login with a confirmed email address before reporting spam
Hi all,
having read many relevant threads and also searched in the comsol manual , i still dont understand the difference (if there is indeed some) between relative error and relative tolerance.
for example:
1)in this thread (also in manual)
www.comsol.com/community/forums/general/thread/13157
''The Relative tolerance value is the relative error tolerance that the ODE solver uses.''
''Use the Relative tolerance edit field to enter a positive number. This number controls how accurately the nonlinear system of equations is solved in each time step. In general, the desired relative error in the solution should be entered here.''
2)in comsol manual
''The software stops the iterations when the relative tolerance exceeds the relative error computed as the weighted Euclidean norm''
suggesting a difference
i need some clarification
V4.4
Thanks in advance
having read many relevant threads and also searched in the comsol manual , i still dont understand the difference (if there is indeed some) between relative error and relative tolerance.
for example:
1)in this thread (also in manual)
www.comsol.com/community/forums/general/thread/13157
''The Relative tolerance value is the relative error tolerance that the ODE solver uses.''
''Use the Relative tolerance edit field to enter a positive number. This number controls how accurately the nonlinear system of equations is solved in each time step. In general, the desired relative error in the solution should be entered here.''
2)in comsol manual
''The software stops the iterations when the relative tolerance exceeds the relative error computed as the weighted Euclidean norm''
suggesting a difference
i need some clarification
V4.4
Thanks in advance
2 Replies Last Post Mar 28, 2016, 4:32 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Djones-
The simple answer is that that relative tolerance and the relative error are two sides of an inequality expression known as the convergence criterion.
Relative error is the calculated error in the solution attempt.
Relative tolerance is the maximum amount of error that you as the user are willing to allow in your solution.
Iterative processes within the solver sequence will continue to iterate on the solution attempt until the calculated relative error approximation drops below the pre-specified relative tolerance.
--
Best regards,
Josh Thomas
AltaSim Technologies
The simple answer is that that relative tolerance and the relative error are two sides of an inequality expression known as the convergence criterion.
Relative error is the calculated error in the solution attempt.
Relative tolerance is the maximum amount of error that you as the user are willing to allow in your solution.
Iterative processes within the solver sequence will continue to iterate on the solution attempt until the calculated relative error approximation drops below the pre-specified relative tolerance.
--
Best regards,
Josh Thomas
AltaSim Technologies
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Thank you Josh