BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:Constructing and Assessing Exact G-Optimal Designs
- Montgomery\, D (Arizona State)
DTSTART;TZID=Europe/London:20110830T140000
DTEND;TZID=Europe/London:20110830T143000
UID:TALK32559AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/32559
DESCRIPTION:Methods for constructing G-optimal designs are rev
iewed. A new and very efficient algorithm for gene
rating near G-optimal designs is introduced\, and
employed to construct designs for second-order mod
els over cuboidal regions. The algorithm involves
the use of Brents minimization algorithm with coor
dinate exchange to create designs for 2 to 5 facto
rs. Designs created using this new method either m
atch or exceed the G-efficiency of previously repo
rted designs. A new graphical tool\, the variance
ratio fraction of design space (VRFDS) plot\, is u
sed for comparison of the prediction variance for
competing designs over a given region of interest.
Using the VRFDS plot to compare G-optimal designs
to I-optimal designs shows that the G-optimal des
igns have higher prediction variance over the vast
majority of the design region. This suggests that
\, for many response surface studies\, I-optimal d
esigns may be superior to G-optimal designs.\n
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:Mustapha Amrani
END:VEVENT
END:VCALENDAR