A MULTIPLE-PARADIGM SYSTEM FOR RANGELAND PEST-MANAGEMENT

Polycultural agroecosystems, such as rangelands, are too complex and p oorly understood to permit precise numerical simulation. Management de cisions that depend on behavioral predictions of such ecosystems there fore require a variety of knowledge sources and reasoning techniques. Our approach to designing a computer system that provides advice conce rning such ecosystems is to incorporate various reasoning paradigms an d apply whatever paradigm is most appropriate to each task arising in the advice process. This approach is based on a particular process des cription of expert human problem solving that uses four different reas oning paradigms: model-based reasoning (MBR); case-based reasoning (CB R); rule-based reasoning (RBR); and statistical reasoning. The process description is implemented in CAse-based Range Management Adviser (CA RMA), a computer system for advising ranchers about the best response to rangeland grasshopper infestations. CARMA attempts to emulate the h uman ability to integrate multiple knowledge sources and reasoning tec hniques in a flexible and opportunistic fashion. The goal of this appr oach is to enable computer systems to optimize the use of the diverse and incomplete knowledge sources and to produce patterns of reasoning that resemble those of human decision-makers.