DECONVOLUTION OF GAS-CHROMATOGRAPHIC PROFILES FOR MULTICOMPONENT MODEL MIXTURES AND ESSENTIAL OILS BY THE SIMPLEX ALGORITHM

The modified simplex algorithm (MDS) was applied to find the correct mixing ratios of multicomponent model solutions and essential oils for reconstructing these mixtures on the basis of their GC profiles. The MDS found the mixing ratio by maximizing the pattern similarity of simulated GC profiles to that of the true mixture. Five aqueous model solutions containing ten volatile compounds in different proportions at mg 1(-1) levels and five natural essential oils were mixed in different ratios to make the target mixtures. When the MDS was applied to standardized GC data for mixtures and model solutions or essential oils, the GC profiles of mixtures were deconvoluted into the correct combinations of model solutions or essential oils. For the essential oils, the ratios estimated from the initial deconvolution by MDS were useful as weighting factors to improve the accuracy of deconvolution. When irrelevant GC data on model solutions or essential oils were added to the data base for deconvolution, their effects were neglected by the MDS assigning very small values to them.