Abstract—Generally, research that dealt with the selection problems for optimization techniques or structures in relational data ware houses supports these problems by considering only a single criterion of optimization. The optimization criteria may be the response time of query execution, the number of inputs/outputs between the main memory and the disk, the space allocated to store the index or materialized views, or the number fragments required by the administrator of the data warehouse when using the fragmentation technique. The present work deals with the problem of selecting the horizontal fragmentation technique while considering both the number of I/O between memory and disk during decisional queries and the number of fragments, as two objective functions to minimize. To reduce the scope of choice solutions, we are based on a scalar method, called compromise method. The method is complemented by the principle of Pareto front to infer the best solutions. The study has been experimented on APB1 benchmark of data warehouse.
Index Terms—Data warehouse, optimization, multiobjective, pareto.
Mohamed Barr is with ESI (Ex. INI), Algiers, Alegria (e-mail: m_barr@esi.dz).
Cite:Mohamed Barr, "Bi-Objective Optimization Based on Compromise Method for Horizontal Fragmentation in Relational Data Warehouses," International Journal of Machine Learning and Computing vol.3, no. 3, pp. 250-254, 2013.