Good morning,
in the very extensive 2018 Manual the cds variable is explained in section 3.4.1 and then on page 50 an example is given. Is there a reason, that during continuation the opt is used, while during extension the cds is used instead? Should the cds variable be declared global in both the .m-file and the ode file?
Thank you, best regards
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The continuation itself is a numerical problem with particular numerical choices such stepsizes and tolerances. When starting a new continuation run, you specify these with "opt", but when you want to extend a run, you (should) want to apply the same settings as before, hence you pass on "cds" which then exists (or call "global cds"). You would better not specify cds in the .m-file that defines your system.
From: lanast lanast@users.sourceforge.net
Sent: Tuesday, March 10, 2020 9:55 AM
To: [matcont:discussion]
Subject: [matcont:discussion] Options "cds" and "contset" with Matcont CL
Good morning,
in the very extensive 2018 Manual the cds variable is explained in section 3.4.1 and then on page 50 an example is given. Is there a reason, that during continuation the opt is used, while during extension the cds is used instead? Should the cds variable be declared global in both the .m-file and the ode file?
Thank you, best regards
Regarding the two options "MaxStepSize" and "MaxCorrIters", is is possible that choosing the maximum step size too small, can lead to the correction not been able to reach the solution (for the given tolerance) in the number of iteration specified ?
So, if we choose a small number of MaxCorrIters, could it be we are not able to land near curve due to the step vectos being short and the continuation algorithm aborts?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
You may want to set MaxNewtonIters=10; (instead of 3=default).
We apply a few real Newton iterations and then use the same Jacobian matrix for the later updates. So that's typically a speedup, unless you want to ensure convergence.
MaxStepsize has nothing to do with finding the new point or not, it's just telling how coarse the bifurcation curve will be sampled.
Cheers, Hil
From: lanast lanast@users.sourceforge.net
Sent: Saturday, March 14, 2020 2:13 PM
To: [matcont:discussion]
Subject: [matcont:discussion] Options "cds" and "contset" with Matcont CL
Regarding the two options "MaxStepSize" and "MaxCorrIters", is is possible that choosing the maximum step size too small, can lead to the correction not been able to reach the solution (for the given tolerance) in the number of iteration specified ?
So, if we choose a small number of MaxCorrIters, could it be we are not able to land near curve due to the step vectos being short and the continuation algorithm aborts?
Thank you Professor Meijer, I was not aware that the MaxNetwonIters is the amount of the steps with a "real" Jacobian, before using the same. This might explain the behaviour of our nonlinear problem.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Good morning,
in the very extensive 2018 Manual the cds variable is explained in section 3.4.1 and then on page 50 an example is given. Is there a reason, that during continuation the opt is used, while during extension the cds is used instead? Should the cds variable be declared global in both the .m-file and the ode file?
Thank you, best regards
The continuation itself is a numerical problem with particular numerical choices such stepsizes and tolerances. When starting a new continuation run, you specify these with "opt", but when you want to extend a run, you (should) want to apply the same settings as before, hence you pass on "cds" which then exists (or call "global cds"). You would better not specify cds in the .m-file that defines your system.
From: lanast lanast@users.sourceforge.net
Sent: Tuesday, March 10, 2020 9:55 AM
To: [matcont:discussion]
Subject: [matcont:discussion] Options "cds" and "contset" with Matcont CL
Good morning,
in the very extensive 2018 Manual the cds variable is explained in section 3.4.1 and then on page 50 an example is given. Is there a reason, that during continuation the opt is used, while during extension the cds is used instead? Should the cds variable be declared global in both the .m-file and the ode file?
Thank you, best regards
Options "cds" and "contset" with Matcont CL
Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/matcont/discussion/762214/
To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/
Regarding the two options "MaxStepSize" and "MaxCorrIters", is is possible that choosing the maximum step size too small, can lead to the correction not been able to reach the solution (for the given tolerance) in the number of iteration specified ?
So, if we choose a small number of MaxCorrIters, could it be we are not able to land near curve due to the step vectos being short and the continuation algorithm aborts?
You may want to set MaxNewtonIters=10; (instead of 3=default).
We apply a few real Newton iterations and then use the same Jacobian matrix for the later updates. So that's typically a speedup, unless you want to ensure convergence.
MaxStepsize has nothing to do with finding the new point or not, it's just telling how coarse the bifurcation curve will be sampled.
Cheers, Hil
From: lanast lanast@users.sourceforge.net
Sent: Saturday, March 14, 2020 2:13 PM
To: [matcont:discussion]
Subject: [matcont:discussion] Options "cds" and "contset" with Matcont CL
Regarding the two options "MaxStepSize" and "MaxCorrIters", is is possible that choosing the maximum step size too small, can lead to the correction not been able to reach the solution (for the given tolerance) in the number of iteration specified ?
So, if we choose a small number of MaxCorrIters, could it be we are not able to land near curve due to the step vectos being short and the continuation algorithm aborts?
Options "cds" and "contset" with Matcont CL
Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/matcont/discussion/762214/
To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/
Thank you Professor Meijer, I was not aware that the MaxNetwonIters is the amount of the steps with a "real" Jacobian, before using the same. This might explain the behaviour of our nonlinear problem.