Re: Does dlyap (control package) works?

[Nicholas Jankowski]

On Wed, Dec 7, 2016 at 5:31 PM, Hernán J. González <address@hidden> wrote:

I'm not sure if I'm missing something stupid.

My issue is with dlyap function from "control" package

https://octave.sourceforge.io/control/function/dlyap.html

 dlyap(C,D)

is supposed to solve

 C*X*C' + D = X

but it does not seem to work for me:

 p=0.3, q=1-p;

 C=[ q p 0 ;  q 0 p ; q 0 0];

 D = [0 0 0; 0 0 0 ; p p p];

X1=dlyap(C,D)                                                                                                          

   0.034394   0.029959   0.023144

   0.029959   0.055090   0.021713

   0.023144   0.021713   0.316853

C*X1*C'+D

 0.034394   0.029959   0.023144

 0.029959   0.055090   0.021713

 0.323144   0.321713   0.316853

See the discrepancy in the last row.

The correct solution seems to be

A =

   0.50518   0.47078   0.36057

   0.53824   0.50828   0.38328

   0.64640   0.62326   0.54754

I'm using :

GNU Octave Version: 4.0.2

Operating System: MINGW32_NT-6.1 Windows 7 Service Pack 1 i686

package control 3.0.0 

_______________________________________________
Help-octave mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/help-octave

I can confirm the same results using Octave 4.2.0 on Windows 7, 64-bit and control 3.0.0

I don't know the purpose of the function, but it definitely falls under the matlab-incompatible category:

Matlab 2016b:

>> p=0.3, q=1-p;
p =
    0.3000
>> C=[ q p 0 ;  q 0 p ; q 0 0];
>> D = [0 0 0; 0 0 0 ; p p p];
>> X1=dlyap(C,D)
X1 =
    0.5052    0.4708    0.3606
    0.5382    0.5083    0.3833
    0.6464    0.6232    0.5475
>> C*X1*C'+D
ans =
    0.5052    0.4708    0.3606
    0.5382    0.5083    0.3833
    0.6464    0.6232    0.5475

unless someone can explain why not, this appears to be something that could be filed as a bug at bugs.octave.org.  Hernan, would you mind doing so?