Genetic programming approach to predict a model acidolysis system

https://doi.org/10.1016/j.engappai.2009.01.010Get rights and content

Abstract

This paper models acidolysis of triolein and palmitic acid under the catalysis of immobilized sn-1,3 specific lipase. A gene-expression programming (GEP), which is an extension to genetic programming (GP)-based model was developed for the prediction of the concentration of major reaction products of this reaction (1-palmitoyl-2,3-oleoyl-glycerol (POO), 1,3-dipalmitoyl-2-oleoyl-glycerol (POP) and triolein (OOO). Substrate ratio (SR), reaction temperature (T) and reaction time (t) were used as input parameters. The predicted models were able to predict the progress of the reactions with a mean standard error (MSE) of less than 1.0 and R of 0.978. Explicit formulation of proposed GEP models was also presented. Considerable good performance was achieved in modelling acidolysis reaction by using GEP. The predictions of proposed GEP models were compared to those of neural network (NN) modelling, and strictly good agreement was observed between the two predictions. Statistics and scatter plots indicate that the new GEP formulations can be an alternative to experimental models.

Introduction

The physical and nutritional properties of fats and oils, hence commercial value, not only depend on their fatty acid composition but also on triacylglycerol (TAG) composition (Iwasaki and Yamane, 2000), since the distribution of fatty acids on the glycerol backbone determines the structure of a TAG. Modification of fats and oils by enzymatic interesterification has long been practiced to improve their functional and nutritional properties (Macrae and Hammond, 1985). Acidolysis is the most commonly used method of interesterification. Lipids are modified by incorporation of fatty acids into specific positions of TAGs by specific or nonspecific lipases (Hoy and Xu, 2001). Production of cocoa butter equivalents from low cost fats (Wang et al., 2006), reduction of saturated fatty acid content of fats (Balcao et al., 1998) and production of structured lipids (Yankah and Akoh, 2000) are well-studied examples.

Acidolysis reactions are widely used for lipid modifications to improve functional and nutritional properties. The efficiency of the acidolysis reaction depends on reaction parameters. Major factors affecting the performance of the reaction are substrate mole ratio, reaction temperature and reaction time (Willis and Marangoni, 2002). Determination of the effect of each parameter on the reaction is required both for quality improvement and for highest economical turnover. For quality improvement and for highest economical turnover the process must be optimized. In spite of several advantages of enzymatic modification of fats and oils, because of complexity of the reactions, an exact mathematical model is not available for process optimization (Xu, 2003). Gene-expression programming (GEP) may serve as a robust approach and it may open a new area for the development of accurate and effective explicit formulation of acidolysis reactions and also for many food-related biotechnology and bioengineering problems.

Genetic Algorithms (GAs) and genetic programming (GP) have been found to offer advantages to deal with system modelling and optimization, especially for complex and nonlinear systems. GP has been applied to a wide range of problems in artificial intelligence, engineering and science, chemical and biological processes, and mechanical models including symbolic regression. In recent years only a few studies have been reported related to the use of GAs in the field of food science. Izadifar and Jahromi (2007) used GAs for the optimization of vegetable oil hydrogenation process, Dutta et al. (2005) studied optimization of a protease production process by using GAs; Hanai et al. (1999) applied GAs for the determination of process orbits in the koji making process and Liu et al. (2007) used GAs to predict moisture content of grain drying process.

An explicit neural network formulation (ENNF) that predicts the production of major reaction products of interesterification of palmitic acid and triolein as a function of experimental parameters; substrate ratio (SR), temperature (T) and time (t), has recently been performed by Çiftçi et al. (2008). However, a GEP-based explicit formulation for enzymatic interesterification, to the best knowledge of the authors, has not yet existed in the literature. Therefore, the purpose of this study is to develop a GEP-based mathematical model for the production of major products of a model interesterification reaction. For this aim, enzymatic interesterification reactions of palmitic acid and triolein were carried out and the amount of major reaction products (1-palmitoyl-2,3-oleoyl-glycero (POO), 1,3-dipalmitoyl-2-oleoyl-glycerol (POP) and triolein (OOO)) were determined. The data taken from experimental study were utilized in training and testing the developed models. The performance of the proposed models was compared to neural networks model developed by Çiftçi et al. (2008).

Section snippets

Materials

During this research, triolein (⩾99%), palmitic acid (⩾98%) and immobilized sn-1,3 specific lipase (Lipozyme IM, immobilized from Mucor miehei, (42 U/g) were used. All solvents used were of HPLC grade.

Enzymatic acidolysis

Triolein (0.1 mM) and palmitic acid (0.2–0.6 mM) were dissolved in 5 mL hexane in 50 mL erlenmayer flasks. Reactions were carried out with 10% enzyme concentration (based on weight on substrates) in a rotary shaking incubator (New Brunswick Scientific, model Nova 40, USA) at 200 rpm, at 40, 50 and 60 °C.

Overview of GP and GEP

GP is a search technique that allows the solution of problems by automatically generating algorithms and expressions. These expressions are coded or represented as a tree structure with its terminals (leaves) and nodes (functions) (Koza, 1992). GP applies GAs to a “population” of programs—typically encoded as tree-structures. Trial programs are evaluated against a “fitness function” and the best solutions selected for modification and re-evaluation. This modification–evaluation cycle is

Results and discussion

The major factors affecting the product formation were investigated in this study. Acidolysis reactions were carried out at different substrate ratios, temperatures and times. The enzyme concentration was chosen as 10% (based on weight of substrates) because of the recommendations for the enzyme concentration for this kind of acidolysis reactions (Xu et al., 1998). Here, SR is the ratio of moles triolein to the moles of palmitic acid. SR range was kept between 0.17 and 0.5. Because its maximum,

Conclusions

In conclusion, this study reports a new and efficient approach for the formulation of an acidolysis reaction using GEP for the first time in the literature. The proposed GEP model is empirical and based on experimental variables obtained from the experimental study. The results of the GEP model are compared to explicit neural network model developed by Çiftçi et al. (2008) with which the results are found in excellent agreement. The model gives a fast and practical way of estimation of reaction

Acknowledgement

The writers are grateful to the Gaziantep University Research Projects Administration Unit for providing support during the research reported in this paper.

Ozan Nazım Çiftçi received his M.S. and Ph.D. in Food Engineering from University of Gaziantep. He is recently working for Ph.D. He is currently a professor at University of Gaziantep. His research interests lie in lipid biotechnology and structured lipids.

References (21)

There are more references available in the full text version of this article.

Cited by (26)

  • A novel predictive model for estimation of cobalt leaching from waste Li-ion batteries: Application of genetic programming for design

    2018, Journal of Environmental Chemical Engineering
    Citation Excerpt :

    A notable advantage of GEP is its capability to specify accurately the phenotype given the sequence of a gene, and vice versa which is assigned as Karva language [10]. The GEP method has been applied to recognize the manner of non-linear systems such as modeling of composite material fatigue behavior [11], prediction and scoring of human protein-protein interactions [12], prediction of discharge coefficient in rectangular side weirs [13], modeling of alumino silicate compressive damage [14], modeling and optimization of As(III) and As(V) adsorption on the tannin-formaldehyde (TFA) and tannin-aniline-formaldehyde (TAFA) resins [15] and the acidolysis behavior modeling of triolein and palmitic acid under the catalysis of immobilized sn-1, 3 lipase [16]. Evolutionary soft computing methods were also applied in the field of lithium-ion battery optimization and leaching reaction analysis.

  • Development and validation of a GEP model to predict the performance and exhaust emission parameters of a CRDI assisted single cylinder diesel engine coupled with EGR

    2015, Applied Energy
    Citation Excerpt :

    Compared to polynomial regression techniques, GEP strategy provides a superior alternative as it precludes the need to conform to predefined fitness functions [35–37]. Given sufficient experimental data during training, GEP has been observed to have the potential in adapting to complex, multidimensional and non-linear relationships accurately [38]. The present study investigates the possibility of harnessing the synergetic benefits of CRDI and EGR operation to explore an in situ option to contain the NOx penalty of CRDI operation while retaining its advantages of lower PM–BSFC characteristics.

  • A comparative study of GEP and an ANN strategy to model engine performance and emission characteristics of a CRDI assisted single cylinder diesel engine under CNG dual-fuel operation

    2014, Journal of Natural Gas Science and Engineering
    Citation Excerpt :

    Moreover, as compared to classical regression techniques, GEP methodology provides a superior alternative as it precludes the need to conform to predefined fitness functions (Nazari, 2012; Nazari and Riahi, 2011). Given sufficient experimental data during training, GEP has been observed to emulate complex, multidimensional and non-linear relationships accurately between the chosen design and response variables (Çiftçi et al., 2009). The present experimental study was undertaken with a perspective to investigate the possibility of exploiting the benefits of low BSFC and soot emission spectra of CRDI operation by minimizing the expected NOx penalty through CNG dual-fuel operation (Roy et al., 2014c).

  • An evolutionary computational approach for formulation of compression index of fine-grained soils

    2014, Engineering Applications of Artificial Intelligence
    Citation Excerpt :

    GP has several advantages over the conventional and other similar techniques. A notable feature of GP and its variants is that they can produce highly nonlinear prediction equations without a need to pre-define the form of the existing relationship (Sette and Boullart, 2001; Javadi et al., 2006; Çiftçi et al., 2009; Guven, 2009; Guven et al., 2009; Gandomi et al., 2010; Gandomi and Alavi, 2011; Rezania et al., 2011; Tsai, 2011; Azamathulla et al., 2011; Alavi et al., 2011; Chen et al., 2012; Azamathulla, 2012; Mahmood et al., 2013). Multi-expression programming (MEP) (Oltean and Dumitrescu, 2002) is new variant of GP.

  • Nonlinear modeling of soil deformation modulus through LGP-based interpretation of pressuremeter test results

    2012, Engineering Applications of Artificial Intelligence
    Citation Excerpt :

    One of the main features of GP is its ability to generate greatly simplified prediction equations without assuming prior form of the existing relationship (Alavi et al., 2011). For the last decade, GA, GP, variants of GP, and other well-known soft computing techniques such as artificial neural networks (ANNs) and support vector machines (SVM) have been pronounced as alternative methods for simulating the behavior of engineering problems (Cheng et al., 2002; Xie et al., 2006; Lin et al., 2006; Chau, 2006; Javadi et al., 2006; Rezania and Javadi, 2007; Muttil and Chau, 2007; Jia et al., 2008; Ciftci et al., 2009; Javadi and Rezania, 2009; Pal et al., 2011; Rezania et al., 2011; Gandomi et al., 2011a; Gandomi and Alavi 2011). LGP (Brameier and Banzhaf, 2007) is a new subset of GP.

View all citing articles on Scopus

Ozan Nazım Çiftçi received his M.S. and Ph.D. in Food Engineering from University of Gaziantep. He is recently working for Ph.D. He is currently a professor at University of Gaziantep. His research interests lie in lipid biotechnology and structured lipids.

Sibel Fadıloğlu received her M.S. and Ph.D. in Food Engineering from University of Gaziantep. She is currently a professor at University of Gaziantep. Her research interests lie in enzyme technology.

Fahrettin Göğüş received his M.S. and Ph.D. in Food Engineering from University of Gaziantep. He is currently a professor at University of Gaziantep. His research interests lie in Maillard reactions, food dehydration, oil and fat technology.

Aytac Guven received his M.S. and Ph.D. in Civil Engineering from University of Gaziantep. He is currently a professor at University of Gaziantep. His research interests lie in applications of artificial intelligence (neural networks, genetic programming, machine learning and fuzzy logic) in engineering problems.

View full text