[Scilab-users] How to solve sparse nonlinear system by scilabe

[Stéphane Mottelet]

Hello,

fsolve cannot handle a sparse jacobian, but if you use the Newton's 
method (less than 5 lines of Scilab code),  the sparsity of the jacobian 
will be taken into account in the linear system resolution step.

S.

Le 11/08/2014 14:29, Michael J McCann a écrit :
> Since the network is passive and hence stable, have you considered 
> treating it as a pseudo-dynamic system (by giving capacity to hold 
> fluid to each sector) and letting it converge to a solution as the 
> flow adjusts itself to equilibrium?  We used to do this sort of trick 
> with analog computers many years ago and electrical resistance 
> networks actually behave that way. Mike.
> On 2014-08-08 7:55, Jonah wrote:
>> Hello,
>>
>> I'm trying to solve a pressure-flow network system which is sparse, 
>> nonliear
>> and huge, around 1k-2k variables. It seems to me fsolve in scilab can't
>> handle such kind of the problem quite efficiently.
>>
>> Do you have good suggestion for solve sparse nonlinear system using 
>> scilab?
>> Thanks a lot.
>>
>> Regards,
>>
>> Jonah
>>
>>
>>
>> -- 
>> View this message in context: 
>> http://mailinglists.scilab.org/How-to-solve-sparse-nonlinear-system-by-scilabe-tp4031012.html
>> Sent from the Scilab users - Mailing Lists Archives mailing list 
>> archive at Nabble.com.
>> _______________________________________________
>> users mailing list
>> users at lists.scilab.org
>> http://lists.scilab.org/mailman/listinfo/users
>>
>