Development of a decision support system for estuarine modelling using the case-based reasoning technology

Computer modelling is regarded as an important investigative tool, in the formulation of correct guidelines and effective environmental policies for estuaries. Its ability in revealing important aspects of the estuarine dynamics under different conditions has provided strong motivation for the use and success of this approach. However, computer models are normally developed for specific cases, which limits their applicability. Integration of numerical modelling and artificial intelligence techniques has the potential for creating advanced computational environments to assist in design of models for coastal hydrodynamic systems. Case based reasoning is a knowledge-based technique, which helps to capture and structure past experience in the form of cases studied. Through the comparison among similar cases CBR systems provide a support for the user in understanding novel situations and solving new problems. CBR for estuarine modelling (CBEM) imparts a way of building, an intelligent computational environment where the available expertise about estuaries and the described models can be stored and reused. Through an intelligent interface, CBEM allows users to define the characteristics required: for the numerical simulation. Based on the user's input the modelling process starts from determining possible correlation between a new phenomenon to be simulated and the previous studies contained in the system. In this paper, the main emphasis is focused on the estuary and model description and, the retrieval of the models to be used in new simulations. Special attention is given to the organisation of categorisation indices adopted from different classifications of estuaries and models and, to the formal criteria established for retrieving "similar"' computational programs for their use in a given problem.