The Research Of Multidimensional Data Model And Multidimensional Computation

Data Warehouse offers data to Decision Support System. On-Line Analytical Processing can analyze multidimensional data of data warehouse. Data model is the core of data warehouse research, but the models presented currently can't express the complex data structure of electric power load, therefore, we can't assemble data-streams based on these models. In order to solve these problems, the dissertation studies data model of data warehouse and multidimensional computation in detail.The dissertation presents an integrated multidimensional data model and defines several data-cube operations based on the model. The model can express the complex data structure of electric power load efficiently. Besides offering general aggregation operations, such as roll-up and drill-down, the model has also several characteristics listed below: different dimensions sharing one bottom level data; offering indirect roll-up operation; having transformation of dimensions and measures.Multidimensional computation is the core of the On-Line Analytical processing, and performance of the computation project mainly depends on two aspects: the first one is aggregation time; the second is computation precision. This dissertation puts forward a multidimensional computation project based on the linear expressions of data-streams. This project can provide fast aggregation of data-streams, with better performance in computation time and storage space than that of previous projects. Experiment shows the scheme can reduce errors.In the end, research results of the dissertation are summarized, after that, some orientations of multidimensional data model and multidimensional computation for further research are also provided.