V. K. Kalichkin, D. S. Fedorov, V. S. Riksen, K. Yu. Maksimovich
Siberian Federal Scientific Centre of Agro-Biotechnologies of the Russian Academy of Sciences, Krasnoobsk work settlement, Novosibirsk region, Russia
E-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Volume 25 No. 12
Date of paper submission: 18.03.2025, date of review: 26.05.2025, date of acceptance: 29.07.2025
Published: 12/31/2025
Abstract. The purpose of the research was to develop crop rotation models and predict their possible productivity based on the application of a genetic algorithm and machine learning using data from long-term field experiments. Methods. Time series of data on productivity of 9 types of crop rotations at three levels of agrochemical application, obtained in the forest-steppe of the Ob region of the Novosibirsk region by the Siberian Research Institute of Agriculture and Chemicalization of Agriculture of the Siberian Federal Scientific Center of the Russian Academy of Sciences during 1999–2019, were used in the research. Scientific novelty. Based on the analysis of historical data from long-term field experiments, a probabilistic matrix of transitions of agricultural crops in crop rotation was constructed. Hybrid crop rotation modeling, which includes the combined use of a genetic algorithm and the XGBoost machine learning model, allowed to establish the regularities of crop alternation in the 16-year cycle of crop rotations and their possible productivity depending on the level of agricultural intensification on the territory of a particular spatial object. Results. According to the results of modeling it is also revealed that at the extensive level it is possible to obtain high productivity of crop rotations without systematic application of fallow, but it is necessary to combine cereals, legumes and annual grasses. At the normal level of intensification, the modeling did not reveal the need to include a fallow field in the crop rotation while maintaining the role of crop rotation and increasing the importance of winter rye. At application of a full complex of chemicalization means it is proved that intensification of farming does not eliminate completely the importance of optimal alternation in the crop rotation of crops diverse in biological properties. The proposed methodology, based on the combined use of the genetic algorithm and XGBoost machine learning, allows to analyze key factors that determine the optimal crop sequence and crop rotation productivity, and to build their best options taking into account the level of application of agricultural technology intensification tools.
Keywords: crop rotation, productivity, modeling, evolutionary algorithm, machine learning
Acknowledgements. The authors express their gratitude to the candidate of agricultural sciences, leading researcher of the fertility laboratory of the Siberian Federal Scientific Center of the Russian Academy of Sciences N. V. Vasilyeva and the candidate of agricultural sciences, head of the Crop Rotation Laboratory of the Siberian Federal Scientific Center of the Russian Academy of Sciences G. M. Zakharov for the information provided on field experiments with crop rotations.
For citation: Kalichkin V. K., Fedorov D. S., Riksen V. S., Maksimovich K. Yu. Modeling of crop rotations using genetic algorithm and machine learning based on field experiment data. Agrarian Bulletin of the Urals. 2025; 25 (12): 2054‒2063. https://doi.org/10.32417/1997-4868-2025-25-12-2054-2063 (In Russ.)
Download the full text of the article
