Research On Materialized Views Selection In Data Warehouse

The technology of data warehouse (DW) is a new branch of database technology. It faces to analytical envirmment and covers the limitation of traditional database system absolutely, providing strongly supporting to enterprises'anlysis and decision. The data warehouse is a repository of information collected from multiple,possibly heterogeneous,autonomous,distributed databases. The information stored at the data warehouse is in form of views,referred to as materialized views. The query responding time can be speeded by pre-storing. This thesis studies the static and dynamic method of materialized view selection, and gives some effective and pracitical algorithm.The selection of the materialized views is one of the most important decisions in designing a data warehouse. Many models are built to study the materialized view selection problem, and the AO View Graph model is one of the most useful models. Different algorithms for the materialized view selection based on AO View Graph model are given. In order to solve it, the definition of AO View Graph model and some important notions and cost models are introduced.Based on the AO View Graph model, the method and strategy using greedy algorithm for selecting materialized views are proposed. We give a greedy algorithm with fixed ratio band and maintenance restriction.The method for materialized views selection by using genetic algorithm is presented. The solution of problem should be converted into a binary string in terms of AO View Graph given. Genetic operation is presented, and fitness function is defined. A genetic algorithm is proposed which can automatically amend invalid genes produced by the process of crossover and mutation.For solving the shortage of the static algorithms, the methods of dynamic management of materialized views are studied. The validity of this algorithm is...