97 Retrieval Model of Slope Stability Evaluation System Based on Cluster Analysis and Genetic Algorithm

Case-based reasoning technique is one of the artificial intelligence methods developed recently. This paper studies the retrieval model of slope stability evaluation system based on Case-based reasoning. Aimed at existent problem of K-Nearest Neighbor strategy (KNN), dynamic cluster method is used to organize index for slope cases, and the cases are classified into different typical sub-base cases according to property or failure style of slope, which could contract case retrieval space and reducing retrieval time. Through analyzing the influence degree on slope stability evaluation result of each factor and its historic data, genetic algorithm combined KNN is adopted to optimize weight, and a rather objective weight value could be denoted for each attribute into increase quality. Practical engineering slopes are applied to test the retrieval model system. And the results show that cluster analysis method could raise retrieval efficiency, and the optimizing calculation by genetic algorithm combined KNN for weight is objective, effective, and simple. And this retrieving model could raise retrieval efficiency and accuracy of slope case stability evaluation system.