The Application Research On Use Of Improved Classification Model In Component Repository Of Job Shop Scheduling

Software reuse is an effective measure to solve software crisis. Component-Based software development (CBSD) is the most potential way of software resue, and component repository is the core of CBSD. Component's classification and the description is the core of component repository.This article takes Dalian youth funding project"Research on Component Repository of Intelligent Optimization Algorithm for Job Shop Scheduling"as a foundation , takes component's classification and the retrieval as subjects. In view of characteristic of small scaled and specialized of Component Repository of Intelligent Optimization Algorithm for Job Shop Scheduling, proposed one kind of improved classification model with dynamic weight and digitized facet based on study of existing component classification and in management.The new model has made two improvements to the faceted classification model:①By introduction of Code - atom terminology comparative table to transforms description sequence of user's retrieval condition and the components in the component Repository into the digital sequence,so that reduced the difficulty of component match when retrieving.②In new model has introduced one dynamic weight to facet which is decided by customer attention.In a component libeary which is classified with faceted classification model,when each facet has the fixed priority,thus completely neglected the preference of user.However the introduction of the dynamic weight can reflect this customer's preference appropriate,therefor raises the the accuracy ratio. When carries out the component retrieve,fistly,organized the digitized description sequence of every facet to an vector,secondly,organized digitized description sequences of every component in the Repository into a matrix,with the help of multiplication that introducted in the article,can transform process of component retrieving to a multiplication of a vector and a matrix. This kind of retrieving has the high efficiency and the high accuracy ratio characteristic, especially qualify small scale domain component Repository.Practice of building component Repository of intelligent optimization algorithm for JSP validates the feasibility of the new classification model.