Materialized View Selection Using Vector Evaluated Genetic Algorithm

Data Warehousing is an approach in which data from multiple heterogeneous and distributed operational systems (OLTP) are extracted, transformed and loaded into a central repository for the purpose of decision making. Since such database stores huge amounts of historical data, it is necessary to devise methods by which complex OLAP queries can be answered as fast as possible. OLAP is an approach which facilitates analytical queries accessing multidimensional databases. Using materialized views as pre-computed results for time-consuming queries is a common method for speeding up analytical queries. However, some constraints do not allow the systems to save all possible views. Therefore, one of the crucial decisions that data warehouse designers need to make is in the selection of the right set of views to be materialized. This paper focuses on solving the multi-objective view selection problem using a Vector Evaluated Genetic Algorithm (VEGA) approach subject to disk space constraint.