Four Application Models for Artificial Neural Network and Genetic Algorithm in Case Based Reasoning Cycle

Case based reasoning(CBR) is a problem solving methodology rather than a technology. This paper mainly studied the application of artificial neural network (ANN) models, i.e. back propagation network (BPN),adaptive resonance theory (ART1), self organization feature mapping (SOFM), and genetic algorithm (GA) from theoretical perspective. For the case retrieval and adaptation in CBR cycle, four application models, which can provide guidance for CBR practical application, were proposed and their algorithms were also presented and discussed respectively. These four models are: ①case clustering based on ART1 or SOFM; ②case similarity calculation based on BPN; ③weighting value optimization for k nearest neighbors (K NN) based on GA; ④case automatic adaptation based on GA and BPN.