Masters Theses
Date of Award
6-1988
Degree Type
Thesis
Degree Name
Master of Science
Major
Planning
Major Professor
George E. Bowen
Committee Members
Kenneth Kenney
Abstract
Expert systems are interactive computer programs that incorporate judgement, expertience, rule of thumb, and "expert opinion." They exhibit characteristics typically associated with intelligent behavior to mimic the way that human experts evaluate information and draw conclusions. In effect, expert systems manipulate knowledge, while conventional programs manipulate data. Expert systems are a fast-developing branch of Artifial Intelligence with broad application. Although expert systems to aid planners are not yet operational, research is under way into applications such as site analysis and zoning implementation.
This thesis reviews expert systems and the areas to which expert systems have been successfully applied, discusses their potential applications/contributions in planning analysis, and explores how useful they may become in land-use/environmental planning through a demonstration project called Site Evaluation Consultant (SEC). SEC is a rule-based system which evaluates the candidate site(s) for a solid waste landfill from a multidisciplinary point of view, given a particular land use situation. The system employs the Personal Consultant Easy expert system shell developed by Texas Instruments Corporation. As a prototype expert system, it performs well within its defined boundaries.
The thesis also describes the process of developing a small expert system, including interviewing experts and capturing and structuring knowledge. It further describes the process of automating experts' knowledge bases and the nature and operation of an "expert system shell." In addition, the thesis discusses the system critique and future extension of SEC.
The major benefits of SEC arise from the advantage that expert systems have both "conventional" computing and manual techniques. With SEC, it is also found that the immediate practical benefit of the system's development is not the resulting computer program, but the explicit structuring of the decision-making process model in a site evaluation for a sanitary landfill. The SEC also has some limitations. In particular, expert systems are more suited to "deep and narrow" domains of expertise, though they do have some application in "wide and shallow" domains. Here they can be useful as aids to refining knowledge. More research needs to be done for the systematic knowledge acquisition methodologies and better "human window" facilities. In general, system designers and users need to become acquainted with what expert system methdology can do and how different tools assist in meeting user requirements. In planning applications, the goal must not be to produce "expert systems" per se; but rather to produce systems in which expert system tools and algorithmic procedures are used for what they do best.
The thesis concludes that (1) the expert system application in planning should focus on aiding the user instead of replacing the experts and (2) the user should keep a critical attitude during a session and remain in control of the decisions made. As a consequence, users would continue developing their expertise, and would not follow the errors of a faulty system. Even though expert systems techniques cannot replace a hard-earned understaning of some phenomena, they, if properly applied, allow planners to spend a larger proportion of their time on creative endeavors and less on technical drudgery.
Recommended Citation
Jung, Hee-Yun, "Expert systems in planning decision support: an application to site evaluation for solid wate landfill. " Master's Thesis, University of Tennessee, 1988.
https://trace.tennessee.edu/utk_gradthes/13244
