PaPer Ital. J. Food Sci., vol. 28 - 2016 121 - Keywords: Jerusalem artichoke, microwave-assisted drying, effective moisture diffusivity, response surface methodology, genetic algorithm - OptimizatiOn Of micrOwave-assisted drying Of Jerusalem artichOkes (HeliantHus tuberosus l.) by respOnse surface methOdOlOgy and genetic algOrithm e. karacabey1,*, c. baltaciOglu2, m. cevik3 and h. kalkan4 1Food Engineering Department, Engineering Faculty, Suleyman Demirel University, Isparta, Turkey 2Department of Food Engineering, Faculty of Engineering, Niğde University, Konya, Turkey 3Agriculture and Fisheries Directorate, Ministry for EU Affairs, Ankara, Turkey 4Computer Engineering Department, Engineering Faculty, Suleyman Demirel University, Isparta, Turkey *Corresponding author: erkankaracabey@sdu.edu.tr AbstrAct the objective of the present study was to investigate microwave-assisted drying of Jerusalem artichoke tubers to determine the effects of the processing conditions. Drying time (Dt) and effec- tive moisture diffusivity (EMD) were determined to evaluate the drying process in terms of dehy- dration performance, whereas the rehydration ratio (rhr) was considered as a significant quality index. A pretreatment of soaking in a Nacl solution was applied before all trials. the output power of the microwave oven, slice thickness and Nacl concentration of the pretreatment solution were the three investigated parameters. the drying process was accelerated by altering the conditions while obtaining a higher quality product. For optimization of the drying process, response sur- face methodology (rsM) and genetic algorithms (GA) were used. Model adequacy was evaluated for each corresponding mathematical expression developed for interested responses by rsM. the residual of the model obtained by GA was compared to that of the rsM model. the GA was suc- cessful in high-performance prediction and produced results similar to those of rsM. the analy- sis and results of the present study show that both rsM and GA models can be used in cohesion to gain insight into the bioprocessing system. mailto:erkankaracabey%40sdu.edu.tr?subject= 122 Ital. J. Food Sci., vol. 28 - 2016 1. INtroDuctIoN the Jerusalem artichoke (Helianthus tubero- sus L.) has been gaining increasing attention due to the potential use of this plant as a feed- stock for the synthesis of new products and the awareness of its significant health benefits. the storage form of carbon in Jerusalem artichokes, inulin, makes this plant attractive compared to the majority of crops that store carbon as starch (KAYs and NottINGHAM, 2008; VAN Loo et al., 1995; WAtHErHousE and cHAttErtoN, 1993). In spite of its high potential usage in the food in- dustry, consumption of this plant as a raw ma- terial is limited due to changes during its post- harvest period (cAbEZAs et al., 2002; MoDLEr et al., 1993; tAKEucHI and NAGAsHIMA, 2011). therefore, increasing the Jerusalem artichoke shelf-life by processing is of prime importance, and dehydration of its tubers should also be considered in this regard. Various drying tech- nologies have been extensively used as a pres- ervation technique in the food industry. specific technologies, such as microwave-assisted dry- ing, for grains, crops and foods have been well documented (AL-HArAHsHEH et al., 2009; GIrI and PrAsAD, 2007; sHArMA and PrAsAD, 2001). the main reasons to consider the use of mi- crowave energy are to accelerate the drying pro- cess, improve product quality, and reduce costs (AL-HArAHsHEH et al., 2009; GIrI and PrAsAD, 2007; McLouGHLIN et al., 2003). However, ad- ditional effort is required to standardize micro- wave technology in the drying process. For this reason, microwave-assisted drying requires in- vestigation in terms of the underlying physical phenomena, such as the mechanism of molecu- lar transfer. Effective moisture diffusivity is one of the parameters used to evaluate the drying of food materials from the point of view of intra- molecular mass transfer, since transfer of water molecules throughout the solid matrix is gener- ally a rate-controlling step in drying processes (DADALI et al., 2007). Another significant step is to optimize processing variables according to desired targets including faster and more effi- cient processing and improved product quality. response surface methodology (rsM) is a sta- tistical procedure frequently used for process optimization. It uses quantitative data from an appropriate experimental design to determine and simultaneously solve multivariate prob- lems. the equations describe the effect of the test variables on the responses, determine in- terrelationships among test variables and repre- sent the combined effect of all test variables in the response. this approach enables an exper- imenter to efficiently explore a process or sys- tem. In recent years, other optimization tech- niques have also been developed and adapted to food processes. In process engineering design, genetic algorithms (GAs) are considered a novel technique (GoLDbErG, 2001). For highly com- plex and nonlinear processes, researchers have reported successful GA applications in analyz- ing the osmotic dehydration of kiwifruit (FAtHI et al., 2011a) and carrot slices (MoHEbbI et al., 2011a), and plant oil extraction from cloves by supercritical co 2 (HAtAMI et al., 2010). to our knowledge, there are no reported stud- ies on the microwave-assisted drying of Jeru- salem artichokes as well as its optimization in terms of drying performance and quality char- acteristics. therefore, the objective of this study was to investigate and optimize the processing conditions of microwave-assisted drying of arti- choke tubers. Additionally, GA was conducted to evaluate its performance in the optimization of the proposed drying technique. 2. MAtErIALs AND MEtHoDs 2.1 Preparation of samples Fresh Jerusalem artichoke tubers were pur- chased from the local market and stored at 4°c. the tubers were peeled and sliced at a specified thickness by using a lab-scale slicer on which the thickness was adjusted in the range of 1-10 mm. All slices had the same projected area (30*40 mm, wide*length) to avoid its effect on drying due to any change; the slice thicknesses for each trial were changed as presented in ta- ble 1. Microwave output power was another pro- cess variable that was examined at three levels (100, 200, and 300 W), as shown in table 1. the third variable was the concentration of the pre- treatment solution. the experimental design was planned such that there were some trials (run order of trials was 1, 9, 15, and 16; table 1) ex- cluding the Nacl in pretreatment, and in the re- maining trials the slices were treated with Nacl solution (table 1) to determine the effect of salt on the drying characteristics of interest. the pre- treatment was carried out with Nacl solutions of specified concentrations (table 1) at 25°c with controlled agitation for a period of 2 h. Af- ter pretreatment, the samples were removed and rinsed with distilled water to remove the solute that had adhered to the surface and then dried in a microwave oven at the output power speci- fied in table 1. In the case of samples that were not subjected to pretreatment, aliquots of 50 g of tuber slices were directly dried in the micro- wave oven (details provided below), whereas pre- treated samples were weighted as 50 g after im- mersion in Nacl solution for 2 h (table 1). the initial moisture content of Jerusalem artichokes was determined by placing the tubers in a con- ventional oven at 105°c until no further change in weight of the sample was observed. the av- erage moisture content of fresh Jerusalem arti- choke tubers was 81.77 ± 0.89%. the moisture content of any pretreated J. artichoke slice did not vary significantly; even with a 2% (w/v) Nacl Ital. J. Food Sci., vol. 28 - 2016 123 concentration in the pretreatment solution. this may be due in part to the low temperature level and short duration of the pretreatments. 2.2 Drying equipment and experimental method A programmable domestic microwave oven (samsung-MW71E, Malaysia) with a maximum output power of 800 W and wavelength of 2,450 MHz was used for drying. Aliquots of 50 g of pretreated or fresh tuber slices were spread on a glass dish (dried and weighed before use) as a single layer and placed on the center of the turntable of the microwave cavity. Drying was performed for each trial at the microwave out- put power levels specified in table 1. Moisture loss was measured periodically (60-s intervals) by taking out and weighing the dish on a digi- tal balance. the drying process continued un- til the desired moisture content was attained (< 10%, w/w). trials were carried out according to the experimental design including the process- ing conditions and run order for each trial (ta- ble 1). the rehydration ratio (rhr) was also de- termined for J. artichoke slices dried according to each trial specified in table 1. the rhr is an important quality parameter to evaluate the dry- ing process in terms of product quality. Dried slices were immersed in warm water (50°c) and their weight gain was monitored until it stabi- lized. the rhr was calculated as a ratio of net weight gain to initial sample amount. 2.3 Theoretical approach to effective moisture diffusivity the effective moisture diffusivity (EMD) was determined to obtain information about the mechanism of moisture transfer and complexi- ty of the drying process. It was defined by Fick’s second law with the assumption that diffusion is the only physical mechanism to control the transfer of water molecules to the surface. Ar- tichoke slices prepared at different thicknesses were assumed to be an infinite slab, since oth- er directions were large enough compared to the thickness. thus, moisture movement was only throughout thickness. Fick’s second law for moisture movement was solved with the fol- lowing assumptions: - the particle was homogenous and isotropic - the material characteristics were constant, and the shrinkage was negligible - mass transfer was in one direction - moisture was initially uniformly distributed throughout the mass of a sample - the pressure variations were negligible - evaporation occurred only at the surface - surface diffusion was ended, so the moisture equilibrium arises on the surface - effective moisture diffusivity was constant ver- sus moisture content during drying - resistance to mass transfer at the surface was negligible compared to the internal resistance of the sample - mass transfer was represented by a diffusion- al mechanism the following analytical solution of Fick’s sec- ond law proposed by crANK (1975) was used to calculate the effective moisture diffusivity. Eq. (1); where D eff is the effective moisture diffusivity (m2s-1), L is the half thickness (drying from both sides) of slab (m), Mr was the fractional mois- table 1 - Experimental design of microwave drying and corresponding responses. Standard order Run order Power Thickness NaCl conc. Drying time Effective diffusivity*10-8 Rehydration ratio (W) (mm) (g/100 mL) (min) (m2/s) 9 1 200 2 0 6 0.79 3.27 16 2 200 4 1 8 1.61 2.76 13 3 200 4 1 9 1.58 3.18 2 4 300 2 1 4 0.77 4.47 14 5 200 4 1 14 1.50 3.71 4 6 300 6 1 5 7.62 3.19 3 7 100 6 1 96 0.42 4.28 15 8 200 4 1 12 1.31 3.18 5 9 100 4 0 75 0.26 3.98 12 10 200 6 2 11 3.37 2.97 7 11 100 4 2 83 0.24 3.57 11 12 200 2 2 5 0.68 5.29 1 13 100 2 1 33 0.11 4.45 8 14 300 4 2 4 4.60 3.57 10 15 200 6 0 9 3.77 4.03 6 16 300 4 0 7 3.32 3.19 124 Ital. J. Food Sci., vol. 28 - 2016 ture ratio, t was the drying time (s). M t was the moisture content of the material at any time, t; M i was the initial moisture content of the mate- rial before drying; and M e was the equilibrium moisture content of a dehydrated artichoke slice, all moisture content values were in dry basis. For long-term drying, only the first term of Eq.(1) was used to explain the drying procedure. the equilibrium moisture content (M e ) was assumed to be zero for microwave-assisted drying. the fi- nal equation to calculate the EMD was as follows: Eq. (2); Further simplification of Eq. (2) resulted in a straight-line equation as Eq. (3); Eq. (3); the effective moisture diffusivity was calculat- ed by fitting Eq. (3) to the curve of ln(Mr) vs. time (Fig. 1), and the results are presented in table 1. 2.4 Experimental design Drying time (Z 1 ), effective moisture diffusivi- ty (Z 2 ), and rehydration ratio (Z 3 ) were the re- sponses used to optimize the process variables by response surface methodology (rsM). A box- behnken design was employed in this regard. Independent process variables (X 1 , X 2 , and X 3 ) were microwave output power, slice thickness, and concentration of the pretreatment solution (Nacl); each was specified at three levels with 16 runs including four replicates at the central point. the ranges and levels of independent var- iables are presented in table 1. Minitab (Minitab 15.1.0.0) was used to analyze the experimental data, which were fitted to a second-order polyno- mial regression model including the coefficients of linear, quadratic and two factors interaction effects. the proposed model was as follows: Eq. (4) where Z was the response of the equation, 0b was the constant coefficient, β i was the linear co- efficient (main effect), β ii was the quadratic coef- ficient, and β ij was the two factors interaction co- efficient. the surfaces of the predicted responses were plotted by sigma Plot (v. 8.02; 2002) (sPss Inc. chicago, IL, usA). the values of R2, adjust- ed-R2, and lack-of-fit of models were evaluated to check the model adequacies. 2.5 Optimization by genetic algorithm the genetic algorithm (GA) is a global search algorithm, which is designed to mimic charles Darwin’s principle of “survival of the fittest” to solve complex optimization problems without falling into local optima (GoLDbErG, 2001; Mo- HEbbI et al., 2011b; MorIMoto, 2006). MAtLAb version 2010b (MathWorks, Inc.) was used to op- timize the interested responses of microwave- assisted drying of Jerusalem artichoke tubers as a function of process conditions by the GA. 3. rEsuLts AND DIscussIoN this study was designed to evaluate micro- wave-assisted drying of Jerusalem artichokes and to optimize the process using response Fig. 1 - Linear relation between ln(Mr) and drying time for the slice thickness of 2 mm treated with 1% Nacl and dried at 100 W output power (■), the slice thickness of 2 mm without treatment and dried at 200 W output power (●), the slice thickness of 4 mm treated with 1% Nacl and dried at 200 W output power (▲) and the fitted proposed model line (—). Ital. J. Food Sci., vol. 28 - 2016 125 surface methodology (rsM) and genetic algo- rithms (GA). Drying of J. artichoke tubers re- sulted in good performance with high quality product in terms of drying time (Dt), effective moisture diffusivity (EMD), and rehydration ratio (rhr). Models developed by rsM and GA displayed similar performances to predict the experimental results determined for each in- terested response. Multiple linear regression analysis of the ex- perimental data yielded second-order polynomial models for predicting Dt, EMD, and rhr. Anal- ysis of variance (ANoVA) was conducted to de- termine significant effects of process variables on each response and to fit second-order poly- nomial models to the experimental data. regres- sion equation coefficients of the proposed mod- els and statistical significance of all main effects calculated for each response were obtained. the effects that were not significant (p > 0.05) were stepped down from models without damaging the model hierarchy (table 2). the ANoVA table also showed that the lack of fit was not significant for all response surface models at a 95% confidence level. on the other hand, R2 and Adj-R2 were cal- culated to check the model adequacy as lack-of- fit > 0.05; R2 ≥ 0.98; and Adj-R2 ≥ 0.94 (table 2). 3.1 Drying time Drying time (Dt) is important because it is an index of the drying performance. A reduc- tion in drying time means less energy require- ment for the process. table 2 shows that both microwave power and slice thickness signifi- cantly affected Dt to decrease the moisture content of slices to less than 10% (p ≤ 0.05), whereas a change in the salt (Nacl) concen- tration of the pretreatment solution was not an important factor (p > 0.05). the microwave- assisted drying process, which reduced the moisture content of Jerusalem artichoke to less than 10%, took 4-96 min varying based on the process variables. the Dt decreased as microwave output power increased due to higher energy transfer for unit process time (Fig. 2). A similar microwave power effect on Dt was reported previously (AL-HArAHsHEH et al., 2009; soYsAL, 2004; suMNu et al., 2005). the favorable influence of output power on Dt may be attributed to the heating mecha- nism of microwave technology causing high in- ternal pressure and concentration gradients, which increases the flow of liquid throughout the food (AL-HArAHsHEH et al., 2009; suMNu et al., 2005; WANG and sHENG, 2006). the sec- ond factor that had a significant effect on Dt values was slice thickness (Fig. 2). However, an increase in Dt is not desirable from an eco- nomical point of view, and there was a positive relationship between slice thickness and Dt (table 2 and Fig. 2). Drying time to decrease moisture content under a target level (< 10%) increased with thicker slices, especially when a low output power was set (Fig. 2). A similar result related to the effect of slice thickness on Dt was obtained by GIrI and PrAsAD (2007) studying the drying kinetics and rehydration characteristics of mushrooms that were pro- cessed in microwaves. table 2 - regression coefficients of predicted models for the investigated responses of microwave assisted drying of J. artichoke. Variablea Coefficient Drying time Effective Moisture Diffusivity Rehydration Ratio β 0 130.625*** e 4.055999* 5.918046*** β 1 -1.285*** -0.02718* -0.01159* c β 2 16.562*** -1.60884** -0.79208** d β 3 0.875ns b - 1.029874* β 11 0.003*** 0.0000339ns 3.22E-05** β 22 - 0.096813ns 0.15846*** β 33 - - 0.114851ns β 12 -0.062*** 0.008183*** -0.00139* β 13 - - 0.001977* β 23 - - -0.38475*** model *** *** *** linear *** ** ** quadratic *** ns *** cross-product *** *** *** R2 0.98 0.98 0.98 Adj-R2 0.97 0.96 0.94 Lack-of-fit 0.101 0.312 0.085 a Polynomial model adjusted by backward elimination at the level of 0.05% with the lack-of-fit test, where β 0 is the con- stant coefficient, β i is the linear coefficient (main effect), β ii is the quadratic coefficient, and β ij is the two factors interaction coefficient. b ns, not significant (p > 0.05); c *, significant at p ≤ 0.05; d **, significant at p ≤ 0.01; e ***, significant at p ≤ 0.001. 126 Ital. J. Food Sci., vol. 28 - 2016 3.2 Effective moisture diffusivity Increasing the effective moisture diffusivi- ty (EMD) is desirable in a microwave-assisted drying process, since this technique is expect- ed to create awareness and an improvement in process performance is one of the novelties. the EMD was calculated and used as an index of the rate of the drying process (table 1). the mass transfer of water molecules in potato ma- trix dried using different techniques has been previously studied. For microwave application on potatoes, the calculated diffusivities were report- ed in the range of 1.91*10-8 m2.s-1 to 3.73*10-8 Fig. 2 - response surface for the effects of power and slice thickness treated with 1% Nacl solution on drying time of Jerusalem artichoke slices. m2.s-1 (McMINN et al., 2003), which were com- parable with EMDs (0.11*10-8 m2.s-1 to 7.62*10- 8 m2.s-1 depending on processing conditions) of Jerusalem artichoke slices dried in a microwave oven. According to the results of the ANoVA of EMD, the output power and slice thickness are two important factors affecting the EMD of the drying process (p ≤ 0.05) (table 2). the EMD remained almost constant with changing slice thickness (2-6 mm) at an output power of 100 W (Fig. 3). similarly, changing the output pow- er (100-300 W) did not significantly affect the EMD of 2-mm thick tubers. However, there was a significant interaction between both factors (microwave output power and thickness) (p ≤ 0.05), and the EMD increased when higher val- ues of slice thickness and output power were se- lected (Fig. 3). DAttA and rAKEsH (2013) re- ported that microwave heating is superior com- pared to conventional heating, since significant internal evaporation inside the microwave-heat- ed material leads to additional mechanisms of moisture transport that enhance moisture loss during heating. thus, an increase in microwave power results in more energy transfer to the food material during drying and as a result more in- ternal evaporation resulting in a higher EMD. 3.3 Rehydration ratio rehydration ratio (rhr) is a widely used qual- ity index for dried products. rehydration values provide information about the changes in phys- ical and chemical properties of a dried sample attributed to drying and treatments preceding dehydration (MAsKAN, 2000). to investigate the effect of drying conditions on final product qual- ity, the rhr of dried tuber slices were deter- mined (table 1). the effects of drying conditions on rhr were analyzed by ANoVA and showed that all processing conditions were effective on the rehydration capacity of microwave-assist- ed dried Jerusalem artichoke slices except for the quadratic term of Nacl concentration of the pretreatment solution (p ≤ 0.05) (table 2). Fig- ures 4, 5, and 6 display the change of rhr with output power, slice thickness, and Nacl con- centration. the rhr of dried samples at an out- put power around 250 W was smaller than that measured for slices dried at any other power level, when tuber slices were dried without pre- treatment. on the other hand, a minimum rhr value was measured for J. artichoke pretreated slices dried at an output power of less than 250 W, and tuber slices dried at 200 W had the low- est rhr when they were treated with the high- est concentration (2%) of Nacl solution (Fig. 4). this negative effect of increasing output pow- er on rhr results from quick sample shrinkage due to rapid water loss depending on the inter- nal temperature. the reason for the change in the effect of high output power with the Nacl concentration of the pretreatment solution may Fig. 3 - response surface for the effects of power and slice thickness on EMD of Jerusalem artichoke slices irrespec- tive of pretreatment. * effective moisture diffusivity Ital. J. Food Sci., vol. 28 - 2016 127 result from partial water loss occurring during pretreatment, although the change in the final moisture content of dried slices pretreated with Nacl solution was not significant compared to the water content of fresh tuber slices (data not shown). In other words, microwave-assisted dry- ing finalized in a shorter period for samples with less moisture content compared to fresh ones. thus, the internal temperature of a sample never reaches to its level seen at drying of the sample without pretreatment, which means less shrink- age and high rhr. these results are consistent with the changes in rhr with microwave pow- er also observed by WANG and XI (2005). slice thickness was another factor that had a sig- nificant effect on rhr values. change in rhr was plotted as a function of slice thickness vs. Nacl concentration and slice thickness vs. out- Fig. 4 - response surface for the effects of power and slice thickness treated with 1% Nacl solution on the rehydration ratio of Jerusalem artichoke slices. Fig. 5 - response surface for the effects of power and con- centration (slice thickness of 4 mm) on the rehydration ra- tio of Jerusalem artichoke slices. put power as shown in Figures 5 and 6, respec- tively. the rhr of the dried products decreased with an increase in slice thickness. the effect of Nacl concentration on this trend was signifi- cant when low slice thickness values were con- ducted (Fig. 5). the rhr increased with increas- ing Nacl concentration of pretreatment solution when thinner slices were analyzed (Fig. 5). A de- crease in rhr was also detected with increas- ing thickness under the effect of power (Fig. 6). thickness effects may result from greater vol- umetric heating, which generates higher pres- sure inside the Jerusalem artichoke tuber, re- sulting in boiling and bubbling of the samples and reduced rhrs of the dried products (WANG and XI, 2005). 3.4 Optimal responses An optimization procedure by rsM was con- ducted for all responses as a function of process- ing conditions. the EMD and rhr were maxi- mized, since higher values of these responses means faster drying and better product quality, respectively. the Dt response was minimized be- cause a short process length is preferred due to economical considerations. As a consequence of the optimization procedures for these three dry- ing characteristics, the following operating con- ditions were found to be optimal: power of 235 W; slice thickness of 5.95 mm; and Nacl con- centration of 0.081. 3.5 Genetic algorithms the GAs were used to select the best sub- set of variables and to build predictive regres- sion models in order to study the relationships between the results obtained from the exper- imental trials (Dt, EMD, rhr) and the pro- Fig. 6 - response surface for the effects thickness and con- centration (power of 200W) on the rehydration ratio of Je- rusalem artichoke slices. 128 Ital. J. Food Sci., vol. 28 - 2016 cess parameters. the coefficients of regres- sion models corresponding to Dt, EMD, and rhr are presented in table 3. the residual is an index of model performance where a small- er residual indicates better prediction perfor- mance. thus, residuals between experimental results and predicted values by rsM and GA are shown in Figures 7-9 for each response. Models produced by GA display a similar per- formance in prediction of EMD and rhr val- ues as those produced by rsM. Figure 7 shows smaller residuals of Dt values predicted by models using rsM than GA. Although a per- formance decrease was seen in the prediction of Dt values by GA, this procedure present- ed in this work can be applied for optimiza- tion in microwave-assisted drying of food ma- terials as a rapid and non-destructive inspec- tion method. GAs have been reported as a nov- table 3 - Model coefficients of proposed second order polynomial modela obtained by GA. Variablea Coefficient Drying time Effective Diffusivity Rehydration Ratio β 0 -0.500 0.522 -0.143 β 1 0.118 0.439 -0.267 β 2 0.035 0.024 0.107 β 3 0.488 0.065 0.190 β 11 0.253 0.065 0.481 β 22 -0.129 0.065 0.024 β 33 -0.335 0.439 -0.309 β 12 0.047 0.024 0.107 β 13 0.076 0.060 -0.600 β 23 -0.500 -0.558 0.600 a Polynomial model , where β 0 is the constant coefficient, β i is the linear coefficient (main effect), β ii is the quadratic co- efficient, and β ij is the two factors interaction coefficient. Independent process variables (X 1 , X 2 , and X 3 ) were microwave power, slice thickness, and concen- tration of pretreatment solution (NaCl). Fig. 7 - residuals between experimental results and predicted responses by rsM (∆) and GA (□) models calculated for each trials of drying time in experimental design. el approach in the osmotic drying of kiwifruit by FAtHI et al. (2011b). similarly, MoHEbbAt et al. (2011) reported genetic algorithms as a method with a high potential for optimization in all food processes. 4. coNcLusIoNs the experimental results and their analysis demonstrate the possibility of using this inno- vative method based on microwave technolo- gy for the drying of Jerusalem artichoke tu- bers. to the best of our knowledge, this is the first study on the microwave-assisted drying of Jerusalem artichoke tubers and optimiza- tion of process parameters using rsM and GA procedures. the results of the present work demonstrate the feasibility of the Dt, EMD, Ital. J. Food Sci., vol. 28 - 2016 129 and rhr determinations for accurate predic- tion. the performance of rsM with respect to r2, adj-r2 and lack-of-fit values was accepta- ble. the GA and rsM methods produced sim- ilar models of performance for microwave-as- sisted drying of artichoke tubers. the analy- sis and results from this present study imply that both rsM and GA models can be used in cohesion to gain complete insight into the bi- oprocessing system. Fig. 8 - residuals between experimental results and predicted responses by rsM (∆) and GA (□) models calculated for each trials of effective moisture diffusivity in experimental design. rEFErENcEs Al-Harahsheh M., Al-Muhtaseb A.H. and Magee t.r.A. 2009. Microwave drying kinetics of tomato pomace: Effect of os- motic dehydration. chemical Engineering and Process- ing 48: 524. cabezas M.J., rabert c., bravo s. and shene c. 2002. Inulin and sugar contents in Helianthus tuberosus and cicho- rium intybus tubers: Effect of postharvest storage tem- perature. Journal of Food science 67: 2860. crank J. 1975. “the Mathematics of Diffusion” 2nd ed. ox- ford university Press, England. Fig. 9 - residuals between experimental results and predicted responses by rsM (∆) and GA (□) models calculated for each trials of rehydration ratio in experimental design. 130 Ital. J. Food Sci., vol. 28 - 2016 Dadalı G., Apar D.K. and Özbek b. 2007. Estimation of Ef- fective Moisture Diffusivity of okra for Microwave Dry- ing. Drying technology 25(9): 1445. Datta A.K. and rakesh V. 2013. Principles of Microwave combination Heating. comprehensive reviews in Food science and Food safety 12: 24. Fathi M., Mohebbi M. and razavi s.M.A. 2011a. Application of Fractal theory for Prediction of shrinkage of Dried Ki- wifruit using Artificial Neural Network and Genetic Algo- rithm. Drying technology, 29:8, 918. Fathi M., Mohebbi M. and razavi s.M.A. 2011b. Effect of osmotic dehydration and air drying on physicochemical properties of dried kiwifruit and modeling of dehydration process using neural network and genetic algorithm. Food and bioprocess technology 4: 1519. Giri s.K. and Prasad s. 2007. Drying kinetics and rehy- dration characteristics of microwave-vacuum and con- vective hot-air dried mushrooms. Journal of Food Engi- neering 78: 512. Goldberg D.E. 2001. Genetic algorithms in search, optimi- zation and machine learning. Pearson Education, sin- gapore. Hatami t., Meireles M.A.A. and Zahedi G. 2010. Mathemat- ical modeling and genetic algorithm optimization of clove oil extraction with supercritical carbon dioxide. the Jour- nal of supercritical Fluids 51: 331. Kays s.L. and Nottingham s.F. 2008. “biology and chem- istry of Jerusalem Artichoke Helianthus tuberosus L.”. crc Press, London. Maskan M. 2000. Microwave/air and microwave finish dry- ing of banana. Journal of Food Engineering 44: 71. Mcloughlin c.M., Mcminn W.A.M. and Magee t.r.A. 2003. Microwave drying of multicomponent powder systems. Drying technology 21: 293. McMinn W.A.M., Khraisheh M.A.M. and Magee t.r.A. 2003. Modelling the mass transfer during convective, micro- wave and combined microwave-convective drying of solid slabs and cylinders. Food research International 36: 977. Modler H.W., Jones H.W. and Mazza G. 1993. observations on long-term storage and processing of Jerusalem ar- tichoke tubers (Helianthus tuberosus). Food chemis- try 48: 279. Mohebbat M., Mohammad-r A.t., Fakhri s. and Mohsen Z.s. 2011. Modeling and optimization of Mass transfer dur- ing osmosis Dehydration of carrot slices by Neural Net- works and Genetic Algorithms. International Journal of Food Engineering: 7:2, DoI: 10.2202/1556-3758.1670. Mohebbi M., Akbarzadeh-t M.r., shahidi F. and Zabihi s.M. 2011a. Modeling and optimization of mass transfer dur- ing osmosis dehydration of carrot slices by neural net- works and genetic algorithms. International Journal of Food Engineering 7(2): DoI:10.2202/1556-3758.1670. Mohebbi M., shahidi F., Fathi M., Ehtiati A. and Noshad M. 2011b. Prediction of moisture content in pre-osmo- sed and ultrasounded dried banana using genetic algo- rithm and neural network. Food and bioproducts Pro- cessing 89(4): 362. Morimoto t. 2006. Genetic algorithm. In “Handbook of Food and bioprocess modeling techniques”. s.s. sablani, M.s. rahman, A.K. Datta, and A.s. Mujumdar (Ed.),, crc Press, New York. sharma G.P. and Prasad s. 2001. Drying of garlic (Allium sativum) cloves by microwave-hot air combination. Jour- nal of Food Engineering 50: 99. soysal Y. 2004. Microwave Drying characteristics of Pars- ley. biosystems Engineering 89(2): 167. sumnu G., turabi E. and oztop M. 2005. Drying of carrots in microwave and halogen lamp–microwave combina- tion ovens. LWt-Food science and technology 38: 549. takeuchi J. and Nagashima t. 2011. Preparation of dried chips from Jerusalem artichoke (Helianthus tuberosus) tubers and analysis of their functional properties. Food chemistry 126: 922. Van Loo J., coussment P., Leenheer L., Hoebregs H. and smits G. 1995. on the presence of inulin and oligofruc- tose as a natural ingredients in the western diet. critical reviews in Food science and Nutrition 6: 525. Wang J. and sheng K. 2006. Far-infrared and microwave dry- ing of peach. LWt-Food science and technology 39: 247. Wang J. and Xi Y.s. 2005. Drying characteristics and dry- ing quality of carrot using a two-stage microwave pro- cess. Journal of Food Engineering 68: 505. Watherhouse M.L. and chatterton N.J. 1993. Glossary of fructan terms, In “science and technology of fructans”. M. suzuki and N.J. chatteron (Ed.),, pp. 2. crc Press, boca raton, FL. Paper Received April 3, 2014 Accepted May 28, 2015