A Cost Model for Dynamic Selection of Aggregates to Materialize in Data Warehouses Roberto Reis Monteiro Neto and Maria Luiza Machado Campos Return to Data Warehouses Abstract The efficient processing of analytical queries in decision support systems highly depends on anticipating demands over aggregated data by precomputing and storing these aggregates in the data warehouse. The careful selection of which aggregates to store may save considerable space and maintenance costs. This paper presents a dynamic approach where aggregates are periodically evaluated using a cost model that considers frequence of use, I/O saves and volume expansion. DiSC'01 Copyright ©2002 ACM Inc.