Research and implementation of the inference engine in the rodent community expert system in the Alashan desert area based on a MapGIS platform

Building on previous research and development of the rodent community expert system in the Alashan desert area, we analyzed the relative quantity of Meriones meridianus, Dipus sagitta, and Phodopus roborovskii in three different types of habitats (reclamation, rotation grazing and overgrazing) as well as the dynamic relationships (both quantitative and non-linear) of eight different plant factors. We utilized the MapGIS platform with eight plant factors and three dominant rodents in three habitats, and a database system composed of population data, a knowledge base system composed of image information of three dominant species of rodents, and a model library system composed of 19 ecological models in three habitats, so as to form an inference engine for the dynamic relationship between the relative number of populations in the rodent community and plant factors. The MapGIS was developed using the programming language VB (Visual Basic) and was designed and applied in the expert system.