Feed Wheat Yield Multifactorial Forecasting

Abstract

The paper proposes a transition from reactive states monitoring of complex agrobiotechnical objects (CABO) to proactive monitoring. It involves either preventive assessment of CABO states (with predetermined monitoring intervals) or predictive (continuous) assessment, analysis, diagnostics, and anticipatory multivariate forecasting. The goal is to detect, localize, and prevent unexpected disruptions in the vital functions of agrobiological and technical elements and subsystems of CABO due to diseases (for agrobiological objects), failures, and malfunctions (for technical objects). A key component of both proactive control and CABO state monitoring is multivariate forecasting, considering inherent delays in the feedback loops of proactive monitoring and control systems. The paper introduces a new concept of multifactorial multi-model adaptive forecasting of CABO state parameters. This concept assumes: first, unification of monitoring information heterogeneous in acquisition methods and presentation; second, application of a multi-model approach to constructing a combined multifactorial model for forecasting CABO state indicators; third, adaptability of the multi-model complex structure for multifactorial adaptive forecasting to the quantity and quality of initial monitoring data and forecasted processes properties. An example of implementing the concept in forecasting fodder wheat yield is provided.