Rule-based management for simulation in agricultural decision support systems

Rule-based management systems can offer the farmer or consultant the opportunity to better approximate current or proposed management options, especially as they relate to dynamic conditions in farm fields and across the whole farm. Most existing attempts at rule-based management within simulation-oriented agricultural decision support systems (DSS) involve limited extension of fixed management dates in response to environmental conditions, or involve rules for implementing limited management events such as fertilizer applications or irrigations. A comprehensive rule-based management system for agricultural DSS was developed that allows simulated management events to occur in response to flexible producer-defined rules and to weather and management induced changes in the soil-crop system over time and space. The system provides a simple, English-based rules language, a rules development editor, and software to parse and interpret these rules and provide linkages to application DSS;software packages such as Great Plains Framework for Agricultural Resource Management (GPFARM). Rules can be quickly and easily developed that cover management activities for individual management units (MUs) or groups of MUs. Feedback to the simulation package at each time step provides for generation of site- and time-specific management events across the application. Tests of the rule-based system on: wheat (Triticum aestivum L.), corn (Zea mays L.), fallow and wheat-fallow crop rotations used in eastern Colorado showed that management events were simulated within the correct time windows and in the proper sequence. Dates for simulated events varied as expected across each rotational cycle as a function of temporal conditions. Additional work is anticipated to allow dynamic calculation of event attributes, capture and implementation of producer time priorities, and a simplified menu system for the rule editor. (C) 1998 Elsevier Science B.V. All rights reserved.