Логіко-лінгвістична модель процесу вирощування хлібопекарських дріжджів

Abstract

LOGIC - LINGUISTIC MODEL OF GROWING BAKER’S YEAST

Y. Samoilenko, V. Tregub

National University of Food Technologies

Modern management theory is applied to a broad class of objects that are developed automation systems with properties of self-organization, adaptation, optimization and intelligence. This applies, above all, complex technological facilities that operate under uncertainties, which include devices for growing yeast. In recent years the development of this industry is to improve processes, allowing yeast to increase quality and productivity devices. The process of baking yeast cultivation requires the maintenance of the main factors of the process according to the chosen technological requirements, deviations of at least one of them reduces the generative activity of yeast, which in turn affects the yield of biomass.

Traditional management techniques do not address fully all the features of growing biotechnological processes baking yeast and require significant resources and aimed at stabilizing the main technical parameters of the process. In turn, they provide fully timely response to a variety of production situations that may be related to changing the quality characteristics of raw materials, a significant level of uncertainty of cultivation.

Using intelligent system allows getting new knowledge; identify causal relationships between factors acting on object control, while not requiring precise knowledge of the mathematical model. To present knowledge of the process of growing yeast using fuzzy logic, which is created based on fuzzy logic-linguistic variables allows us to develop methods and algorithms for modeling complex process under uncertainty and incomplete information. Using a neural network allows you to turn the information we need necessarily to synthesize and intelligent process control yeast cultivation.

To build  neural-fuzzy network yeast cultivation process an internal subsystem environment Matlab - subsystem design neural - fuzzy structures ANFIS. ANFIS - an abbreviation Adaptive-Network-Based Fuzzy Inference System - adaptive network fuzzy inference with a single output and multiple inputs that are fuzzy linguistic variables, which can automatically synthesize the experimental data neural-fuzzy network, visualize the structure, if necessary, make changes in its parameters, perform her studies in which the deviation between the results of fuzzy modeling and experimental data is minimal.