Dear A. H. Fraser,
This is a quick reply - I will also ask a little around - Ian
Postlethwaite is really the person to ask.
The issue with poorly conditioned matrices or ill-posed Riccatti systems
may not necessarily be a problem in terms of the result - the experts
on this are Chiang and Safonov who wrote the Robust Control toolbox.
There should be no problem in applying the factorization to
non-minimum phase systems.
I am not sure where the problems come from. You could try a low-order and
see how this changes the results. If there results are about the same then
I would assume things were OK. You could also try using the mu toolbox
(see page 379 in the book) and see if you get the same result.
Best regards,
Sigurd Skogestad
Received on Thu Sep 12 09:22:21 1996
This archive was generated by hypermail 2.1.8 : Tue Jan 20 2004 - 09:32:14 CET