Controller synthesis with uncertainty

Dear Steve Fisher,

Thank you for your nice comments about our book.

>On p. 327 you state that the multiplicative uncertainty description would be
>less suitable for controller synthesis,

This is not quite what is meant. The multiplicative uncertainty description
in itself is OK - the problem is that there is no explicit penalty on
the input usage - this should come in addition to the uncertainty.

Best regards,
Sigurd Skogestad

>>From smf198@psu.edu Fri Jan 22 21:26:03 1999
>
>Professor Skogestad,
>
> I recently purchased your book, Multivariable Feedback Control, and woul
>d
>like to offer my compliments as well as pose a question. The book is the most
>practical, easy to read text on the subject I have yet to encounter. I am
>currently addressing the problem of representing an input uncertainty. On
>p. 327 you state that the multiplicative uncertainty description would be
>less suitable for controller synthesis, which makes sense after examining
>the augmented plant. Would an inverse multiplicative representation of the
>uncertainty completely resolve the problem? Would it still be suitable for
>robustness analysis?
>Thank you for your time and help.
>
>Respectfully,
>
>Steve Fisher
>
>Pennsylvania State University
>smf198@psu.edu
Received on Mon Jan 25 09:42:29 1999