SYBIL DSS : LOCALIZATION OF AGRICULTURAL RISK ASSESSMENT MODELS

EU Project SYBIL implements a decision support system (DSS) to help farmers intelligently manage crops (reducing environmental impact and increasing economic returns). Agro-meteorological computer models are used to assess the risk of the crop to pest and fungus damage. In particular, we focus on grape and apple models that help farmers predict when diseases and funguses will attack their plants, so they can make intelligent decisions on preventing these attacks. To achieve this goal in an intelligent manner, a hybrid methodology for localizing these models to specific regions has been designed and implemented. Model localization is frequently necessary because models developed in one region often do not produce valid results when used in a different region. The main component of this intelligent localization methodology is a genetic algorithm (GA), an artificial intelligence (AI) search technique. By linking a genetic algorithm to an agricultural risk assessment model, the model become more robust because it is able to adapt to the region in which the model is being used. Preliminary testing indicates this localization methodology has the ability to allow regional agricultural models to be effectively utilized in other regions.