Genetic Algorithms (GAs) in the Role of Intelligent Regional Adaptation Agents for Agricultural Decision Support Systems

Abstract

An AI adaptation methodology designed to assist in transporting agricultural models between regions is presented. A methodology to perform model adaptation (viz. localization) is frequently necessary when models are transported because models developed in one region often do not produce valid results when used in a different region. In this methodology, a GA plays the role of the adaptation agent. By linking a GA to an agricultural model, the model become more robust because it is able to adapt to the region in which the model is being used. This methodology has been implemented within a decision support system (DSS) developed within an EC project called SYBIL. The DSS helps farmers predict when diseases and funguses will attack their plants, so they can make intelligent decisions on preventing these attacks. Preliminary testing within this environment indicates this adaptation methodology has the ability to allow agricultural models developed in one area to be effectively utilized in other regions.