A modified statistical iteration technique for the global optimization of multivariates in chromatographic separations

Summary

A computer approach to the global optimization of multivariates in chromatographic separations is presented using a modified statistical iteration technique, and the optimization of GLC separations is used as the example. In principle this approach should be applicable not only to GC but also to HPLC if the concrete optimization models with respect to the optimized varibles are available. In order to minimize the computation time, an empirical equation related to peak width at half height was also studied. The results have shown that this empirical equation is statistically acceptable and the calculated values of peak widths at half height were good enough when compared with the experimental values with few exceptions within a given region; and that the optimization technique described here is efficient for locating the global optima of the variables as a practical computer approach though it is short of the mathematical perfection.