Graph analytics of the performance of local biotechnical systems in animal husbandry

Abstract

The development of biotechnical systems in animal husbandry is characterized by the level of its intellectual and digital transformation based on neural network technologies and artifi cial intelligence systems. They are supposed to effi ciently recognize and take into account the refl exes, individual and group motivation of animal behavior and implement them in the corresponding local technological subsystems. Pathways to resting and self-service places for animals can be represented in the form of an oriented graph. To analyze it, the authors propose an integral ST-criterion (path-time), characterizing the duration of movement along the graph edges (a logistical infrastructure of the barn) between its vertices. The proposal includes some points of animal service in the corresponding local biotechnological systems (LBTS) (those of milking, feeding, watering, etc.). The graph analytics for each animal helps estimate idle travels from resting to service places (self-service), duration of service in the respective LBTS (working movement), veterinary treatment in case of diseases, total motor activity, total rest time in the box (at least 14 hours for highly productive animals), including duration of night and day rest, abnormal breaks between milkings (over 14 hours) in case of “voluntary” milking in automatic systems (robots), etc. By comparing photo-chronometer indicators with productivity and physiological state of each animal separately and analyzing possible time losses (downtime) at service (self-service) places, deviations in animal behavior, we can get a clear picture of the organization and effi ciency of technological processes on the farm, possible productivity losses and production costs. The graph analytics of local biotechnical systems in cattle breeding equipped with animal identifi cation and video surveillance systems wi ll enable farmers to optimize the on-farm control of technological processes.