The sequential simplex method and multicriteria optimization in analytical chemistry
Öberg, T.
Lecture at the Eastern Analytical Symposium in Somerset, New Jersey, USA, November 14-19, 1999.
Abstract
Optimization of analytical methods usually involves a number of experimental factors and a number of responses. Two optimization methods that have been applied with particular success the last three decades are the basic and the modified simplex methods. Successful applications can be found in most areas of spectroscopy and chromatography.
"Desirability" or "fuzzy set membership" functions is a proven methodology for combining and optimizing several responses simultaneously, e.g. separation time and resolution. The simplex and fuzzy set approaches can then be combined to achieve rapid and simultaneous optimization of several responses. This combined simplex/fuzzy set approach will be compared with classical experimental design methods, e.g. response surface methodology and overlay plots.
Some analytical chemistry applications described in the literature will be discussed together with some more recent laboratory studies. Finally, a short résumé will be given of a web project to popularize the use of systematic experimental optimization methods.
|
Swedish homepage