MIST (Model Identification and Selection Tool): A Knowledge-Based System for Selection of Water Quality Models

A Knowledge-based system, MIST (Model Identification and Selection Tool), is developed for selection of water quality models. A knowledge base is set up using information collected from a number of manuals, literature searches, to facilitate this selection process. The knowledge is then codified into a knowledge-base that can be used within a knowledge-based system. A knowledge-base skeleton, M.4, is used as a backbone to develop the system. The knowledgebase skeleton provides a mechanism to develop rules in production rule format, and includes forward and backward chaining goal seeking mechanisms, and allows the user to provide confidence/certainty factors to the answers provided in the process of consultation. The front-end of the decision support system is developed using integrated development environment (IDE) under visual basic (VB) programming to provide a graphical interface for the end user. MIST consists of a number of modules (knowledge-bases) that are relevant to few classes of modeling environments. Results suggest that MIST can be used as a model selection tool for inexperienced modelers.