@Rouben Rostamian  

Thanks for all, and especially @Rouben Rostamian  . Your code is Ok:).

g:= (x,y)-> `if`(y=sqrt(1-x^2), 100, `if`(x=sqrt(1-y^2), 100,`if`(x=0, 0, 'procname'(args))));

with

Here is our solution: from the code:
x, y, w := algo_12_1([-1,1],[0,1],[6,3],1e-10,100):
Error, (in algo_12_1) unable to store '.2500000000*g(-1., HFloat(0.6666666666))+.2500000000*g(HFloat(-0.6666666667), 1.)' when datatype=float[8]
but this is not well :(.

Regards

@sideshow 

This is not as well for more conditons.:

g:= (x,y)-> `if`(x=0.5, 200*y, 'procname'(args)) or  (y=0.5, 200*x, 'procname'(args));

regards.

@Carl Love 

Thank you for all,

To insert more than one condition?, for example four conditions for the same function g:

g(0,y):=0, g(x,0):=0

g(x,0.5)=200*x ; g(0.5,y)=200*y

Thank you so much.

@acer 

I'm sorry, I do not understand you. I sent the file in which I am working. How do I make the succession of matrix?.

example.mw

@acer @Preben Alsholm

Thank you for your comments. I have a sequence of matrices 5x5, and have to work with the elements of each matrix. Given a matrix A [k] 5x5 , with k: = 0, I have to work with the same elements as follows:

@Kitonum Thanks!!!

@Mac Dude 

Hi,

I resolved this with w[n].

Regards.

@tomleslie 

Thanks.

Best Regards.

@Adri van der Meer , This is correct.

@Carl Love 

when I run the command num_n (1) for example, the system returns x =1; u (1) = ..., u '(1) = ....

The question is:
Through num_n (), How I can extract only u (x)?

Regards.

With eval does not made.

Is there any way to extract results dsolve but only for u (x)?

@Carl Love