Decision Support for Mildew Management in Winter Wheat

Abstract A prototype to a decision support system (DSS) for powdery mildew management in winter wheat is presented. Recommendations of the DSS are based on reasoning in a mathematical model. In order to handle the uncertainty involved in predicting the consequences of a treatment strategy the model is a Bayesian network. A recommendation concerns the treatment of the current mildew problem of a particular field, but to optimize the dosage and timing, future possible treatments are also included in the computations. The optimal treatment strategy consists of the set of mildew treatments with maximal expected outcome, given all current knowledge.