A bi-factorial experimental design was considered to assess moisture variation of sweet potato-quinoa-kiwicha flakes (SP-Q-K) caused by the changes in the rotational speed and steam pressure of a rotary drum dryer (RDD). As it is a design with discrete variables, there is a limitation in the modeling and optimization thus techniques of Artificial Intelligence (AI): Artificial Neural Networks (ANN), Fuzzy Logic (FL) and Genetic Algorithms (GA), were applied, and their prediction ability evaluated. Due to the limitation of data for proper training, the ANN did not allow a correct prediction of the experimental data. Response Surface Methodology (RSM) was employed to obtain the relational equation among the experimental variables, which was used as the objective function with GA, and this allowed moisture optimization. Because of this, it is recommended to integrate RSM and GA into optimization studies. In this research the use of FL among variables, enabled us to get the best prediction adjustment of experimental values (R 2 = 0.99), with a mean absolute error of 0.6±0.66 %, setting a pressure value of 5 atm and a speed value of 6 rpm for flakes at 4.99 % humidity.

ResumenEl objetivo de este trabajo fue comparar la optimización del proceso de extracción de glucosinolatos totales de harina de maca (Lepidium meyenii) (EGTHM) utilizando SR por diseño Box-Behnken (SRBB) con el de AG, en función a x1: temperatura (ºC), x2: etanol (%), x3: relación solvente/materia prima y x4: tiempo de extracción (min). Se identificó y cuantificó los GT utilizando HPLC. Se evaluaron las variables (x1, x2, x3, x4) que influyen en la extracción utilizando un SRBB con el software Statistica y Wolfram Mathematica para los AG. Del desarrollo del SRBB se obtuvo una ecuación de segundo orden con R 2 = 0,74794, p = 1,88248E-10 << 0,05 con error absoluto medio de 11%; lo que indicó la consistencia del modelo. No fue posible obtener un valor óptimo de la EGTHM utilizando SRBB, por la existencia de dos zonas óptimas debido a la configuración de una superficie tipo silla. Empleando AG se obtuvo después de 2000 iteraciones el valor máximo de la función de 17,0986 μmol de GT/g de HM, el cual se alcanzó con 69,9783 ºC, 70,9540% de etanol, relación solvente/materia prima de 10,0488 en 90 min, lo que demuestra la aplicabilidad de los AG. Palabras clave: algoritmos genéticos, superficie de respuesta, optimización, secado, maca (Lepidium meyenii). AbstractThe aim this work was to compare the extraction process optimization of total glucosinolates of maca flour (Lepidium meyenii) (ETGMF) using RS for Box-Behnken (RSBB) Design with that of GA, according to x1: temperature (°C), x2: ethanol (%), x3: ratio solvent/raw material and x4: extraction time (min). TG were identified and quantified using HPLC. The variables (x1, x2, x3, x4) that influence their extraction were evaluated using a RSBB with the software Statistica and Wolfram Mathematica for the AG. From the development of the RSBB, a second order equation with R 2 = 0.74794, p = 1.88248E-10 << 0.05 with 11% average absolute error was obtained; it showed the consistency of the model. It was not possible to obtain an optimal value of the ETGMF using RSBB because of the existence of two optimal zones due to the configuration of a chair surface. After 2000 iterations using GA, the maximum value of the function of 17.0986 μmol of TG/g of MF was obtained, which was reached with 69.9783 °C, 70.9540 ethanol%, 10.0488 ratio solvent/raw material in 90 min, which demonstrates the applicability of the GA.