Abstract:
Expert systems are computer programs designed to mimic human experts in solving problems. This report investigates the feasibility of using expert systems to aid decision-makers in coastal zone management and development planning who are undertaking environmental assessments of coastal resort development proposals. The approach adopted in the study is as follows: First, determine what characteristics an expert system should have in order to make it suitable to address problems often encountered in coastal zone management and development planning. Second, determine what problem-solving methods may be suitably applied using an expert systems approach, in order to derive acceptable solutions to these problems. Third, select a narrow domain from each of three important problem areas for modelling purposes. Identify sources of expertise for each domain and extract the required knowledge from them. Encode this knowledge with the aid of a suitable expert system development tool, so as to form three prototype expert systems. Each expert system is to determine one of the following: (a) The likelihood of occurrence of significant negative effects of septic tank effluent disposal (b) The likelihood of exceeding the recreational carrying capacity for craft on a limited surface area of water (c) The need for different recreational management options for craft using a surface water feature also used by birds Fourth, construct a fourth prototype expert system whose task is to co-ordinate the others and to use their results to help the user (decision-maker) to determine suitable approaches to finding acceptable development constraints and management options for coastal resort development proposals. One major finding is that rule-based expert systems are suitable for addressing the kinds of problems often encountered in coastal zone management and development planning, provided they incorporate certain features. These features are, inter alia: (1) The ability to address multiple goals at a time. (2) The ability to combine forward and backward reasoning in different ways. (3) The ability to call up and communicate with other programs, including other expert systems. (4) The ability to work with mathematical variables and formulae (within the rules). (5) The ability to work with facts and beliefs. Another major finding is that it is feasible to use a rule-based expert system to co-ordinate others and to use their results to help the user (a multi-disciplinary expert) to determine acceptable solutions to multi-disciplinary problems in coastal zone management and development planning.
