Development of an Expert System for Environmental Management Planning: Soil and Water Conservation

A computer-based expert system, EMP-Ex, has been developed for environmental management planning with regard to soil and water conservation, particularly during earthwork activities of development projects. It is a rule based expert system programmed in wxCLIPS 1.62. The EMP-Ex is able to predict soil erosion rate and peak runoff caused by development activities and then recommends suitable strategies for conserving soil and water resources, monitoring the effectiveness of mitigation measures and water quality. The system provides users a checklist of required items and suggestions for steps to be taken in the preparation of Environmental Management Plan (EMP) including planning for emergencies and in the production of an EMP report. The system's knowledge base comprises descriptive and prescriptive knowledge elicited from domain experts and additional supporting information acquired from literature. The domain knowledge was incorporated into the system in the form of production rules that can be updated and referred to through the system. EMP-Ex has been verified and validated to evaluate system capabilities by wxCLIPS facilities, face validation, Turing test, and field study. The results show that EMP-Ex is able to function as good as human experts with a ninety-five percent of confidence level. Through interfacing with other external programmes (e.g. AutoCAD, IDRISI, Microsoft Office, Netscape Navigator etc.), the system extends its capability in sharing and storing raw knowledge and external databases for further reference or updating and provides users the convenience in using additional facilities of the external programmes interfaced. By automating EMP processes, not only can EMP-Ex help EMP planners to improve the quality and the quantity of work, but it can also assist the authorised agencies such as the Department of Environment (DOE) in auditing and revising the plans, especially in the situation of insufficient human experts.

Text FK_2000_49.pdf Download (1MB)

Actions (login required)

View Item