Paper Info
Title
Materialized view selection using evolutionary algorithm for speeding up big data query processing.
Abstract
For speeding up query processing on Big Data, frequent sub-queries or views may be materialized such that the query processing cost is minimized with optimum cost of maintaining the materialized views and/or queries. Materializing frequent sub-queries and views means that resultant data set of the views reside in the memory of one or more nodes in the cluster, so that it reduces the MapReduce cost, submission and scheduling cost of Distributed File System jobs for query processing. We have defined materialized views as resultant data of frequent sub-queries and aggregation functions of a set of Big Data warehousing queries that are saved for enhancing query performance. The problem is defined as a multi-objective optimization problem for minimizing the total query processing MapReduce cost, MapReduce cost for maintaining the materialized views and the number of views selected for materializing with maximized total size of the views selected. We applied Differential Evolution algorithm and NSGA-II to study their performances for developing a recommendation system for selecting views for materializing in Big Data warehousing.
Year
DOI
Venue
2017
https://doi.org/10.1007/s10844-017-0455-6
J. Intell. Inf. Syst.
Keywords
Field
DocType
Big data warehouse,Differential evolution algorithm,Hadoop,Hive,Materialized view,Multi-objective optimization,NSGA-II
Recommender system,Distributed File System,Data mining,Evolutionary algorithm,Scheduling (computing),Computer science,View,Multi-objective optimization,Materialized view,Big data,Database
Journal
Volume
Issue
ISSN
49
3
0925-9902
Citations 
PageRank 
References 
1
0.35
26
Authors
3
Name
Order
Citations
PageRank
Rajib Goswami141.06
Dhruba K. Bhattacharyya222627.72
Malayananda Dutta310.35