Generalized simulated annealing for calibration sample selection from an existing set and orthogonalization of undesigned experiments

Abstract

Generalized simulated annealing (GSA) is an optimization procedure for locating the global optimum (maximum or minimum) of multidimenisonal continuous functions. GSA has been modified for optimization of discrete functions. Selection of calibration samples from an existing set defines discrete optimization and GSA is used to select optimal sets of calibration samples for specific analysis samples. The procedure is applied to near‐infrared spectra. When compared to using the complete set of 37 calibration samples, concentration prediction errors were reduced 50%–100% by using select sets of two to seven calibration samples. Additionally, GSA was able to improve a poorly designed experiment. GSA devised augmented experimental designs such that the overall experimental design (original plus augmented) was more orthogonal than the original.

Chemometrics tools in QSAR/QSPR studies: A historical perspective

“…The mathematical details of GSA could be found elsewhere [54,107,240,241] Here again, after removing a variable in each iteration step, a cost function is computed and is compared to the previous step. Removing a non-significant variable (descriptor) should not lead to obtaining a solution worse than the previous one.…”

“…For reaching this purpose, logical selection of control parameter is necessary. Some rules about the selection of β and C 0 could be found in literature [240,243].…”

Análise multicomponente simultânea por espectrofotometria de absorção molecular UV-VIS

“…Kalivas et al 59 apresentam os fundamentos desse método e sugerem o seu uso no planejamento da série de calibração para análise quantitativa. O generalized simulated annealing também se aplica à otimização de funções discretas, como a seleção de misturas de calibração a partir de uma série já existente 60 . Este aspecto da otimização tem sido objeto de estudos mais recentes como o de Ferré e Rius 61 , segundo o qual uma série de calibração definida por um planejamento fatorial pode ainda ser reduzida usando-se o planejamento Dotimizado, de modo que a sub-série escolhida é aquela que resulta na menor variância dos coeficientes de regressão.…”

Computer‐generated multicomponent calibration designs for optimal analysis sample predictions

“…Simplex has been used as a method for optimizing data acquisition schedule^'^ but can easily converge to local optima. l 6 Once the final design has been determined by VSGSA, the calibration samples are constructed followed by calibration of the model and ultimately concentration predictions of the analysis sample.…”