Re: problem with residue and repeated complex roots

[Ben Abbott]

ok, I checked my local Octave 2.9.17 (which I have some trouble  
with), and I checked Matlab 7.3

-------------------------------------
octave:2> n = 1;
octave:3> d = [1 8 42 104 169];
octave:4> roots(d)
ans =

  -2.0000 + 3.0000i
  -2.0000 - 3.0000i
  -2.0000 + 3.0000i
  -2.0000 - 3.0000i

octave:5> [a b c] = residue(n,d)
a =

  3.9416e-18 - 9.2593e-03i
  -2.7778e-02 - 5.6372e-10i
  -2.3130e-18 + 9.2593e-03i
  -2.7778e-02 + 5.6372e-10i

b =

  -2.0000 + 3.0000i
  -2.0000 + 3.0000i
  -2.0000 - 3.0000i
  -2.0000 - 3.0000i

c = [](0x0)
-------------------------------------

For Matlab 7.3 on Mac OSX I get (essentially) the same result.

-------------------------------------
n = 1;
>> d = [1 8 42 104 169];
>> roots(d)

ans =

 -2.000000031975913 + 3.000000067504366i
 -2.000000031975913 - 3.000000067504366i
 -1.999999968024083 + 2.999999932495634i
 -1.999999968024083 - 2.999999932495634i

>> [a,b,c] = residue(n,d)

a =

 -0.000000000000000 - 0.009259259259259i
 -0.027777777777778 + 0.000000000000000i
 -0.000000000000000 + 0.009259259259259i
 -0.027777777777778 - 0.000000000000000i


b =

 -1.999999999999998 + 3.000000000000000i
 -1.999999999999998 + 3.000000000000000i
 -1.999999999999998 - 3.000000000000000i
 -1.999999999999998 - 3.000000000000000i


c =

     []
-------------------------------------

My results agree with your Octave results, but not with your Matlab.  
What version of Matlab are you using.

Ben

p.s. sorry for not looking at the detail earlier


On Nov 27, 2007, at 8:42 PM, Doug Stewart wrote:

> I downloaded from
>
> http://velveeta.che.wisc.edu/cgi-bin/cvsweb.cgi/~checkout~/octave/ 
> scripts/polynomial/residue.m?rev=1.35;content-type=text%2Fplain
>
> and it does use mpoles.m
>
> but I still get
>
> > > [a b c]=residue(n,d)
>
> ver = 1.35
>
> a =
>
> -9.3613e-018 - 9.2593e-003i
>
> -2.7778e-002 - 5.3894e-010i
>
> 2.3130e-018 + 9.2593e-003i
>
> -2.7778e-002 + 5.3894e-010i
>
> b =
>
> -2.0000 + 3.0000i
>
> -2.0000 + 3.0000i
>
> -2.0000 - 3.0000i
>
> -2.0000 - 3.0000i
>
> c = [](0x0)
>
> To prove that I was using the new file and not an older file I  
> added a line to print out the ver. number.
>
>
> Please test this on your machine.
>
>
> Doug Stewart
>
>
>
> Doug Stewart wrote:
> > Thanks
> > Doug
> >
> > Ben Abbott wrote:
> >
> > > Doug,
> > >
> > > This problem was addressed with a change that made it into  
> > > 2.9.16. If you check the cvs you can replace your residue.m with  
> > > the one there. You'll need mpoles.m as well.
> > >
> > > I should also mention that another bug/feature was fixed just  
> > > yesterday.
> > >
> > > Ben
> > >
> > > On Nov 27, 2007, at 5:11 PM, Doug Stewart wrote:
> > >
> > >
> > > > I am running qtoctave on windows XP octave version 2.9.15
> > > > residue doesn't give the correct answer for repeated complex roots
> > > >
> > > > her is Matlab's result
> > > >
> > > > » n
> > > > n =
> > > >      1
> > > > » d
> > > > d =
> > > >      1     8    42   104   169
> > > > » roots(d)
> > > > ans =
> > > >   -2.0000+ 3.0000i
> > > >   -2.0000- 3.0000i
> > > >   -2.0000+ 3.0000i
> > > >   -2.0000- 3.0000i
> > > > » [a b c]=residue(n,d)
> > > > a =
> > > >   1.0e+005 *
> > > >    1.2662+ 1.8232i
> > > >    1.2662- 1.8232i
> > > >   -1.2662- 1.8232i
> > > >   -1.2662+ 1.8232i
> > > > b =
> > > >   -2.0000+ 3.0000i
> > > >   -2.0000- 3.0000i
> > > >   -2.0000+ 3.0000i
> > > >   -2.0000- 3.0000i
> > > > c =
> > > >      []
> > > >
> > > > and qtoctave's results
> > > >
> > > >
> > > > > > n
> > > > n = 1
> > > >
> > > >
> > > > > > d
> > > > d =
> > > >
> > > > 1 8 42 104 169
> > > >
> > > >
> > > > > > roots(d)
> > > > ans =
> > > >
> > > > -2.0000 + 3.0000i
> > > >
> > > > -2.0000 - 3.0000i
> > > >
> > > > -2.0000 + 3.0000i
> > > >
> > > > -2.0000 - 3.0000i
> > > >
> > > >
> > > > > > [a b c]=residue(n,d)
> > > > a =
> > > >
> > > > -9.3613e-018 - 9.2593e-003i
> > > >
> > > > -2.7778e-002 - 5.3894e-010i
> > > >
> > > > 2.3130e-018 + 9.2593e-003i
> > > >
> > > > -2.7778e-002 + 5.3894e-010i
> > > >
> > > > b =
> > > >
> > > > -2.0000 + 3.0000i
> > > >
> > > > -2.0000 + 3.0000i
> > > >
> > > > -2.0000 - 3.0000i
> > > >
> > > > -2.0000 - 3.0000i
> > > >
> > > > c = [](0x0)
> > > >
> > > >
> > > >
> > > > I haven't looked into why this is happening yet.
> > > >
> > > > Doug Stewart
> > > >
> > > > _______________________________________________
> > > > Help-octave mailing list
> > > > address@hidden
> > > > https://www.cae.wisc.edu/mailman/listinfo/help-octave
> > _______________________________________________
> > Help-octave mailing list
> > address@hidden
> > https://www.cae.wisc.edu/mailman/listinfo/help-octave