Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #4467 > unrolled thread

Shooting method

Started byJoao Pereira <joaopereira9@gmail.com>
First post2011-08-19 10:35 +0000
Last post2011-08-22 10:05 +0000
Articles 5 — 4 participants

Back to article view | Back to comp.soft-sys.math.mathematica


Contents

  Shooting method Joao Pereira <joaopereira9@gmail.com> - 2011-08-19 10:35 +0000
    Re: Shooting method Alois Steindl <Alois.Steindl@tuwien.ac.at> - 2011-08-20 10:16 +0000
    Re: Shooting method Michael Weyrauch <michael.weyrauch@gmx.de> - 2011-08-20 10:18 +0000
    Re: Shooting method Alois Steindl <Alois.Steindl@tuwien.ac.at> - 2011-08-20 10:16 +0000
      Re: Shooting method Frank K <fkampas@gmail.com> - 2011-08-22 10:05 +0000

#4467 — Shooting method

FromJoao Pereira <joaopereira9@gmail.com>
Date2011-08-19 10:35 +0000
SubjectShooting method
Message-ID<j2lea0$nfl$1@smc.vnet.net>
Hi, I'm fairly new to Mathematica and I'm trying to solve a problem using the shooting method. I wonder if you could help me?

I have a system of 6 nonlinear ODE of order 1, three of them have initial conditions, the other three have terminal conditions. I'm trying to use the shooting method to guess the 3 missing initial consitions. The problem I have is that I get several error messages regarding inderteminate expression/ComplexInfinity and Power::1/0 encountered. I also get a message of non numerical value found for a derivative at the intial period.

I am not specifying guesses for the starting conditions, but I think Mathematica automatically does that.

Bottom line, I do not know if the errors I'm finding are a result of a misuse of the shooting method, or if it is something wrong in the way the problem is specified.

Best regards

Joao Pereira

[toc] | [next] | [standalone]


#4482

FromAlois Steindl <Alois.Steindl@tuwien.ac.at>
Date2011-08-20 10:16 +0000
Message-ID<j2o1h9$4q5$1@smc.vnet.net>
In reply to#4467
Hello,
it is possible to specify initial guesses to the shooting method and you
should really do it.
Nonlinear differential equations can exhibit quite nasty behaviour - in
your example it might happen, the the solutions grow to infinity in your
integration interval. Even if the solution starting at the prescribed
initial guess remains bounded, the next iterated could run off.
Also the shooting method is often a source of trouble, especially if the
system possesses unstable trajectories; so it is usually very important
to provide a very good initial approximation to get any solution at all.

Alois

-- 
Alois Steindl,                           Tel.: +43 (1) 58801 / 32558      
Inst. for Mechanics and Mechatronics     Fax.: +43 (1) 58801 / 32598
Vienna University of Technology,         A-1040 Wiedner Hauptstr. 8-10   

[toc] | [prev] | [next] | [standalone]


#4485

FromMichael Weyrauch <michael.weyrauch@gmx.de>
Date2011-08-20 10:18 +0000
Message-ID<j2o1m3$4t2$1@smc.vnet.net>
In reply to#4467
Well, the shooting method implemented in NDSolve works pretty nicely
for me. But without your program (eventually appropriately simplified to 
the essentials) it is hard to help you here.

Michael

Am 19.8.2011 12:35, schrieb Joao Pereira:
> Hi, I'm fairly new to Mathematica and I'm trying to solve a problem using the shooting method. I wonder if you could help me?
>
> I have a system of 6 nonlinear ODE of order 1, three of them have initial conditions, the other three have terminal conditions. I'm trying to use the shooting method to guess the 3 missing initial consitions. The problem I have is that I get several error messages regarding inderteminate expression/ComplexInfinity and Power::1/0 encountered. I also get a message of non numerical value found for a derivative at the intial period.
>
> I am not specifying guesses for the starting conditions, but I think Mathematica automatically does that.
>
> Bottom line, I do not know if the errors I'm finding are a result of a misuse of the shooting method, or if it is something wrong in the way the problem is specified.
>
> Best regards
>
> Joao Pereira
>

[toc] | [prev] | [next] | [standalone]


#4499

FromAlois Steindl <Alois.Steindl@tuwien.ac.at>
Date2011-08-20 10:16 +0000
Message-ID<j2o1i8$4qk$1@smc.vnet.net>
In reply to#4467
Hello,
just a second remark:
some decades ago there existed a very powerful boundary value problem
solver PASVAR (in Fortran, of course) by Lentini and Pereira.
If you are willing to work with Fortran, many excellent codes are
available for that task, especially COLSYS, COLNEW, MULSHOT and
BOUNDSCO. 
Alois
 

[toc] | [prev] | [next] | [standalone]


#4519

FromFrank K <fkampas@gmail.com>
Date2011-08-22 10:05 +0000
Message-ID<j2t9m3$m0q$1@smc.vnet.net>
In reply to#4499
On Aug 20, 6:16 am, Alois Steindl <Alois.Stei...@tuwien.ac.at> wrote:
> Hello,
> just a second remark:
> some decades ago there existed a very powerful boundary value problem
> solver PASVAR (in Fortran, of course) by Lentini and Pereira.
> If you are willing to work with Fortran, many excellent codes are
> available for that task, especially COLSYS, COLNEW, MULSHOT and
> BOUNDSCO.
> Alois

A better approach might be solving the problem iteratively using quasi-
linearization, since you'd be solving a linear boundary value problem
at each step.

[toc] | [prev] | [standalone]


Back to top | Article view | comp.soft-sys.math.mathematica


csiph-web