Gıda üretiminde hammadde karışım değerlerinin matematiksel modelleme ile bulunması = Finding raw material mixture values in food production by mathematical modeling

Abstract

Gıda ve içecek üretimi için kullanılan formüller genellikle on veya daha fazla bileşenden oluşmaktadır. Ürünler; temel ham içeriklere ve baharatlara ek olarak vitamin, renklendirici, lezzet arttırıcı veya koruyucu içerebilir. Bunlardan bir kısmı az miktarda kullanılırken bazı kısımlarında ise işlem yapmak veya dozajlamak çeşitli durumlarda güç olabilir. Reçete geliştirmenin bu iki yönü; çok sayıda içerik bulunması, reçete karmaşıklığı formülasyonlarda sıklıkla karşılaşılan en büyük zorluklardandır. Gıda teknolojilerinde "Deneme-Yanılma" yöntemi yaygın olarak başvurulan bir yaklaşım olmuştur. Deneme yanılma yönteminde, ürün geliştiriciler önceden seçilmiş bileşenlerle bir formülasyon tasarlar ve ardından formülasyondaki bileşenleri belirli spesifikasyonları karşılayacak şekilde yeniden ayarlar. Bu işlem, istenilen ürün elde edilmesine kadar tekrar edilir. Ürün geliştiren kişi, ürünü geliştirme sırasında önceden tanımlanmış bazı özelliklerden ödün verebilmektedir. Böyle bir yaklaşım ise zaman ve maliyet açısından fazlasıyla dezavantajlıdır. Bu nedenle belirli kısıtlarla matematiksel yöntem doğrultusunda gıda ürünü geliştirmek gerekmektedir. Her ne kadar beslenme içeriğini optimize etmek amaçlı Matematiksel Programlama çalışmaları literatürde yer alsa da ürün formülasyonlarının geliştirilmesinde yeterli sayıda çalışma bulunmadığı tespit edilmiştir. Bu çalışmada yeni ürün geliştirme sürecinde ürün formülasyonunun belirlenmesi maksatlı literatürde daha önce kullanılmadığı tespit edilen matematiksel modellemenin kullanılması hedeflenmiştir. Meyve barı ve granolanın hammaddelerinde bulunan temel besin öğeleri (enerji, yağ, protein, KH) ve tekstürleri (sertlik, yapışkanlık, gevreklik) ürünlerin kıstılarını belirlemektedir. Ürünlerin tekstürleri önem derecelerine ve ağırlıklarına göre matematiksel modellerden AHP metodu ile değerlendirilmektedir. SuperDecisions programı AHP'nin seçeneklerini etkili bir şekilde sıralamak, muhakeme yapmak, verileri birleştirmek ve sonuçlarını tahmin etmek için çalışmada kullanılan bir programdır. LINGO programı ise matematiksel modelde kısıt belirlemek için yardımcı olmaktadır. Meyve barı ve granola ürünlerinin üretiminde formülasyon belirlenmesi için iki farklı matematiksel model geliştirilerek en az maliyetle tüketici beklentilerini karşılayacak sonuç elde edilmektedir.

Formulas used for food and beverage production usually consist of ten or more ingredients. Products; may contain vitamins, colours, flavor enhancers or preservatives in addition to basic raw ingredients and spices. While some of the parts are consumed in very small amounts, some of the parts can be difficult to process or dose in various situations. These two aspects of recipe development, namely a large number of ingredients or recipe complexity and the variety of amounts used, are at the root of the most frequently encountered formulation challenges. In food technologies, "Trial and Error" has been a widely used approach. In the trial-and-error method, product developers design a formulation with pre-selected ingredients and then re-adjust the ingredients in the formulation to meet certain specifications. This process is repeated until the desired product is obtained. The product developer may compromise some predefined features during product development. Such an approach is extremely disadvantageous in terms of time and cost. This thesis study started with the examination of the raw materials and ingredients used in the production of fruit bars and granola products from food products. It enables the determination of optimum raw material mixing ratios by considering energy, nutrients and texture, which are indispensable constraints during the production phase. A food bill of materials is the stage of preparing the raw materials of the product. This process is carried out by bringing together all component raw materials in the right proportions in accordance with a certain size and quantity. The proportions of the raw materials mixed in the recipe are determined during the development processes. In this study, a method is developed by using a mathematical model instead of repeated experiments to obtain the desired properties (taste, color, shelf life, etc.) in the food product. Although Mathematical Programming studies to optimize nutritional content are included in the literature, it has been determined that there are not enough studies in the development of product formulations. This study, it is aimed to use mathematical modeling, which has not been done before in the literature, to determine the product formulation in the new product development process. The importance degrees and weights of the texture (hardness, stickiness and brittleness) that are valuable in food production, as well as nutritional elements such as energy, oil, protein, KH, of the raw materials were obtained by the AHP method and used in the creation of the mathematical model. SuperDecisions is a program used to combine reasoning and data to effectively rank AHP's options and predict their consequences. The LINGO program, on the other hand, helped to determine the constraint in the mathematical model. As a result, two different mathematical models were developed to determine the formulation of two different food products, and optimum raw material mixture values were obtained at the least cost and to meet consumer expectations. Most of the work in the food sector is focused on increasing the product range and improving the existing recipes. New ingredients are tested that can maximize taste, consistency, colour, appearance and flavour. Meanwhile, food manufacturers pay attention to not exceeding their production budgets, optimizing shelf life, reducing time-to-market and meeting regulatory requirements. Food formulation; It is the business of creating food and beverage products that meet the taste, consistency, flavor, nutritional and many other expectations of consumers and is a constantly evolving subject. Formulas used for food and beverage production usually consist of 10 or more ingredients. Products may contain vitamins, colours, flavor enhancers or preservatives in addition to basic raw ingredients and spices. While some of them are used in very small amounts, it may be difficult to treat or dose the other part. These two aspects of recipe development, namely the large number of ingredients or the complexity of the recipe and the variety of quantities used, underlie the biggest constant formulation challenges. The aim of this thesis is to get rid of the traditionally followed repetitive tests in food production. For this purpose, two food products (fruit bar and granola) were selected and the raw materials and contents in the formation of the products were examined. In addition to these, texture properties (hardness, brittleness, stickiness) were taken into account. The problem of determining the raw material mixing ratios at the optimum level is emphasized. For this purpose, two different mathematical models have been developed. With the solution of the developed models, new constraints were added to the literature. In the model established for food formulation, an overview of the strategy and analysis method that guarantees fast and accurate formulation is presented. Optimization has been made in formulation with food manufacturers' recipe accuracy and product consistency, ideal taste and consistency, minimum cost and number of experiments. As a result, this thesis study has been provided to guide food manufacturers in their product development processes.