Improvement of cereal harvest programming methods using computer simulation information technology

Abstract

The paper is dedicated to a topical scientific and applied problem – the development of information technology of computer modelling intended for programming the yield of agricultural crops. The paper describes information technology of computer modelling of the yield of agricultural crops (on the example of Avena sativa subsp. nudisativa), which is based on the application of Bayesian methods to modelling and prediction in conditions of statistical, parametric and structural uncertainty. The study was based on the materials of laboratory experiments carried out in conditions close to natural, on the prediction of physiological processes occurring in plants under the influence of regulated and unregulated factors. Proposed approach described the change in the productivity of grain crops, in particular Avena sativa subsp. nudisativa, depending on the parameters of plant growth and development, photosynthetic apparatus and duration of its functioning. Scientific novelty of the work was application of probabilistic and statistical models in the form of Bayesian networks in the system of programming the yield of agricultural crops. The paper considered several scenarios of the combined effect of growth regulators and herbicides on the productivity of Avena sativa subsp. nudisativa. Net productivity of photosynthesis was chosen as the target variable of the studied process. Mathematical models in the form of Bayesian network turned out to be adequate for the process chosen for modelling. Achieved error of model classification was about 20%. The model structure was built in Genie 2.0 modelling system. It was found that by researching and simulating potential opportunities of ecological features of plants, it was possible to achieve an increase in yield by reducing the doses of herbicides and growth regulators by their combined use, which significantly increased the crop quality. Proposed information technology uses methods of intelligent data analysis, has a modular structure and can be used separately and as part of other information and analytical systems.