Research On The Decision Support System Based On Model Driven Architecture

Decision Support Systems is mainly constituted by the Data warehouse and decision reasoning, and with results showing these compose a complete decision process. As a basic method of inductive inference in the decision process, OLAP (on-line analysis process) plays a very important role. And now, the modeling and design of OLAP depend on the specific structure of table and data warehouse of the business. The achievement of modeling and design of OLAP is hard to reuse, a new modeling and design of OLAP needs to re-analyze and re-design.This thesis descirbs a development framework of decision support system based on the MDA. By using UML profile and CWM (common ware metamodel) models PIM (platform independent model) and PSM (platform specific model) of OLAP respectively. Presenting a rule to descirbs PIM transforms to PSM by using Relation language of QVT (Query/View/ Transformation). And implement this rule in the open-source tool—mediniQVT. Moreover, implements transformation from PSM to SQL under EMF (Eclipse Modeling Framework) is which presented by IBM. Thereby, to realize a model-driven development process based on the model. In this way, the design of OLAP can be specified in the early stages of a DW project, thus reducing the development time and cost, and provide support to DSS (Decision Support Systems). And apply this method on a DSS of an iron and steel enterprise. It indicates that the method in the paper increase the efficiency of System Development compares to the traditional method. Therefore, makes the model having more normative, realizes the share between models. Meanwhile, it profits the re-use and maintain of model.