The Research Of Component Retrieval Algorithm Based On Tree Matching And Genetic Algorithm

Component Based on Software Development (CBSD) is the most effective method for software reuse. The component repository system plays an important role in the process of Reusable Component Based on Software Development. In this system, component classification scheme and retrieval mechanism is one of the basic and kernel functions. An effective mechanism of component retrieval can reduce the cost of software reuse evidently. Therefore, how to retrieve a mass of components effectively in CBSD becomes an important problem.The background, current research status and theories of component classification scheme and retrieval teclinology is discussed in this thesis, then the deficiencies of component retrieval method based facet classification schema are pointed out at present. In point of the features of faceted classification and with the reference of the tree matching model, a new method of component retrieval based on XML and Tree Matching Model is proposed. This method can give dual attention to the recall and precision of component retrieval and has higher retrieval efficiency. At the same time, because of some severe limitations of traditional component retrieval algorithms based on facet, on the basis of Generic Algorithm of Artificial Intelligence, an improved intelligent Component Retrieval Model based on Genetic algorithm with Facet-Weight Self-learning is proposed in this paper. Genetic algorithm based facet weight self-learning algorithm can change the facet weight dynamically in order to improve retrieval accuracy, and risk minimization-based component sampling algorithm is used to solve the insufficiency of training data.Finally, in dot NET environment, the prototype of a component retrieval system is realized, and kernel function modules are introduced. The efficiency and feasibility is proved in the experiment.