NEURAL-NETWORK ARCHITECTURES FOR MONITORING AND SIMULATING CHANGES INFOREST RESOURCE-MANAGEMENT

The ability to include both qualitative and quantitative data makes th e neural network approach a very flexible and powerful tool for modeli ng complex biophysical interactions for resource planning applications . Spatially referenced geographical data from the Hoosier National For est, Pleasant Run Unit, were used to test the neural net approach for simulating changes in objectives based on predicted management activit ies. While the use of GIS provided a higher resolution mapping of fore st resources than was traditionally performed, the added benefit of mo re refined management solutions was obtained. This, coupled with the,a ccurate neural net representation of expert-based management, provide an effective method for automatically examining large variable sets fo r more accurate prediction and simulation of potential change in fores t landscape management.