Demydov V. Decision support systems in the automated control system of technological process of fats with demanded properties synthesis

Research object: process of complex interesterificated fats composition which meet established requirements from a known set of components. Aim of research: decision support system construction which allows to solve a problem of analysis and synthesis by insignificant information volume about structure and property of composite mixes. Research methods: methods of functional, cluster and regress analysis, neural networks modeling. Theoretical and practical results: methods functional "structure-property" construction under the conditions of small sample are advanced. Techniques of fat's fusion temperatures estimation on its structure and of fats with the demanded fusion temperature from component set receptions are received. Scientific novelty: mathematical model of a class "structure-property" in the form of approximating polynoms which due to arguments composition allows to lower essentially necessary volume of training sample under condition of preservation of demanded model accuracy has received thefurther development; for the first time the decision support system for the a fats compounding drawing up problem decision with demanded properties is developed. Implementation degree: results have received practical adoption in a branch scientific research institute at synthesis of fats with demanded properties. Application area: synthesis of fats at the enterprises and development of new kinds of fats in scientific research institutes.