Post Reply 
eigenval / EIGENVALUES
06-02-2015, 09:50 PM
Post: #1
eigenval / EIGENVALUES
I'm calculating eigenvalues of a matrix, [[2,-1,0],[-1,2,-1],[0,-1,2]].
In CAS with eigenvals() I get √2+2, 2, 2-√2, ok
With EIGENVAL() I get "Low accuracy" warning, then Bad Argument Value (like in Home)...

What's the problem?
Thank you

Salvo

∫aL√0mic (IT9CLU) :: HP Prime 50g 41CX 71b 42s 39s 35s 12C 15C - DM42, DM41X - WP34s Prime Soft. Lib
Visit this user's website Find all posts by this user
Quote this message in a reply
06-03-2015, 01:49 AM
Post: #2
RE: eigenval / EIGENVALUES
(06-02-2015 09:50 PM)salvomic Wrote:  I'm calculating eigenvalues of a matrix, [[2,-1,0],[-1,2,-1],[0,-1,2]].
In CAS with eigenvals() I get √2+2, 2, 2-√2, ok
With EIGENVAL() I get "Low accuracy" warning, then Bad Argument Value (like in Home)...

Salvo

I get the same error Sad !!!
Find all posts by this user
Quote this message in a reply
06-03-2015, 06:24 PM
Post: #3
RE: eigenval / EIGENVALUES
Very interesting, you have found an example of matrix that enter an infinite loop when applying Francis numeric algorithm with 2 real shifts (the loop is infinite because the matrix is already in Hessenberg form and the shifts are exact integers). Introducing a small rounding error breaks the infinite loop, I'm updating the giac source code to do that.
Find all posts by this user
Quote this message in a reply
06-03-2015, 06:58 PM
Post: #4
RE: eigenval / EIGENVALUES
(06-03-2015 06:24 PM)parisse Wrote:  Very interesting, you have found an example of matrix that enter an infinite loop when applying Francis numeric algorithm with 2 real shifts (the loop is infinite because the matrix is already in Hessenberg form and the shifts are exact integers). Introducing a small rounding error breaks the infinite loop, I'm updating the giac source code to do that.

thank you!

∫aL√0mic (IT9CLU) :: HP Prime 50g 41CX 71b 42s 39s 35s 12C 15C - DM42, DM41X - WP34s Prime Soft. Lib
Visit this user's website Find all posts by this user
Quote this message in a reply
06-04-2015, 08:37 PM
Post: #5
RE: eigenval / EIGENVALUES
(06-02-2015 09:50 PM)salvomic Wrote:  I'm calculating eigenvalues of a matrix, [[2,-1,0],[-1,2,-1],[0,-1,2]].
In CAS with eigenvals() I get √2+2, 2, 2-√2, ok ...
.............
Salvo

Last week I did an academic year project to one student just on this subject. All counted in detail, I used HP-50G only for the check of final result. DET, EGV, EGVL, REF, rref etc work well there (some results were necessary to finish manually them for the teacher). On HP-50 (Home+CAS+RPN in one mode) it was made much more conveniently than on Prime-emulator. I admire your enthusiasm.
It would be interesting to know in more details, what problem you work hard so?
Find all posts by this user
Quote this message in a reply
06-04-2015, 09:06 PM
Post: #6
RE: eigenval / EIGENVALUES
(06-04-2015 08:37 PM)Hlib Wrote:  ...I admire your enthusiasm.
It would be interesting to know in more details, what problem you work hard so?
thank you for admiration Smile
In this case no particular problem (I like much linear algebra); I've also HP 50g (wonderful), but I aim to understand the potentiality and the limits of the Prime with linear algebra.
In the mean time I'm trying to implement something that Prime (for now) doesn't offer in simple way: LDU (and list of pivots), LDLt factorization and few (very few!) other things the Prime already don't have Smile

∫aL√0mic (IT9CLU) :: HP Prime 50g 41CX 71b 42s 39s 35s 12C 15C - DM42, DM41X - WP34s Prime Soft. Lib
Visit this user's website Find all posts by this user
Quote this message in a reply
06-04-2015, 10:08 PM
Post: #7
RE: eigenval / EIGENVALUES
(06-04-2015 09:06 PM)salvomic Wrote:  ... In the mean time I'm trying to implement something that Prime (for now) doesn't offer in simple way: LDU (and list of pivots), LDLt factorization and few (very few!) other things the Prime already don't have Smile

I understand and do appreciate!
Find all posts by this user
Quote this message in a reply
Post Reply 




User(s) browsing this thread: 1 Guest(s)