Expert systems in forestry: Utilizing information and expertise for decision making

Abstract

Expert systems are computer programs designed to simulate the problem-solving ability of human experts in specialized fields. They incorporate a knowledge base that contains the scientific knowledge and experience for solving specific types of problems, a data base for the facts pertinent to particular problems, and a control program that constructs lines of reasoning to solve the problems. They have been successfully applied in many fields, but very few forestry applications have been reported. The authors have developed an expert system, named PREDICT, to help foresters diagnose pest problems in red pine ( Pinus resinosa, Ait.) stands, based on symptoms easily observed in the field. The program recognizes 28 different causes of symptoms in red pine, and incorporates over 400 inference rules in its knowledge base. The manner in which expert systems solve problems is illustrated with an example from the pest diagnosis system. Preliminary experience with the pest diagnosis application suggests that expert systems can be very effective at solving forestry problems of this type. Other areas of forestry that might benefit from the application of expert systems include silviculture, site evaluation, reforestation, and inventory design. The authors encourage other forestry researchers to investigate possible applications of expert systems. Several topics are discussed that are of importance to anyone considering the application of expert systems to areas of forestry. Tools for developing expert systems are readily available, and they can be implemented on desk-top computers for easy access by professionals in all sectors of forestry. Expert systems provide a framework for presenting the latest scientific knowledge and decision-making expertise in a form that can be readily applied by foresters in the field.