Samuel,
Thanks for your answer. Simply being the determinant small doesn't seem
a problem when using floating point. Probably the bigger problem is the
presence, somewhere, of very dissimile orders of magnitude.
One curious detail in my example is that the solution of the system when
adding these lines (input vector and multiplication by the inverse matrix):
V1 = 1;
B = [V1/R1, V1/Z2, 0, 0]';
phi = inv(A) * B;
the denominator of the resulting node potentials in vector phi should be
of degree 8 but they happen to be just scalars (degree 0) as if the
coefficients were below %eps
Regards,
Federico
On 09/03/2020 09:36, sgoug...@free.fr wrote:
Hello Federico,
cond() and rcond() do not accept rationals, but the determinant of A is very
close to 0:
--> det(A)
ans =
3.573D-10
------------------------
0.0000027s² +7.433D-10s³
This likely explains that the inversion is not reliable.
Regards
Samuel
_______________________________________________
users mailing list
users@lists.scilab.org
http://lists.scilab.org/mailman/listinfo/users
_______________________________________________
users mailing list
users@lists.scilab.org
http://lists.scilab.org/mailman/listinfo/users
