Open AccessDevelopment of a neural network architecture for controlling a technological feeding line in a pigsty
Author(s) -
D A Soloviev,
S M Bakirov,
V A Strelnikov,
O V Logacheva,
D A Kolganov
Publication year - 2022
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/1010/1/012161
Subject(s) - architecture , artificial neural network , production (economics) , relevance (law) , profitability index , computer science , control (management) , quality (philosophy) , object (grammar) , distribution (mathematics) , factor (programming language) , line (geometry) , biology , artificial intelligence , mathematics , business , geography , mathematical analysis , philosophy , geometry , archaeology , finance , epistemology , political science , law , economics , macroeconomics , programming language
The article provides the relevance and description of the development of the architecture of the neural network for feeding controlling in a pigsty. The role of the human factor in systems of the type “man – machine – biological object” is described. In pig breeding, the determining factor is feeding, the profitability of meat production depends on the quality of the diet and distribution modes. The data of the average daily weight gain of a pig aged from 0 to 180 days without violations and with violations of the diet and feed distribution regimes are shown on the example of Russian and European pigsties. The feeding line and the main equipment for harvesting, preparation and distribution of feed are presented. The principle of feeding controlling in pigsties without human participation is described. The architecture of a neural network for feeding control in a pigsty is proposed, which contains 17 main neurons and 15 input values. It also provides a graphical description of the connections of neurons among themselves.