Authors: Amit Kumar; Rahul Singh; T.V. Vijay Kumar
Addresses: School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi-110067, India ' School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi-110067, India ' School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi-110067, India
Abstract: Processing of distributed queries in an efficient manner entails generating query plans that would minimise the total query processing cost. The number of possible query plans increases exponentially with an increase in the number of relations accessed by the distributed query, as well as the number of sites where these relations reside. Consequently, it becomes infeasible to explore all possible query plans. In this paper, this problem has been addressed as a bi-objective optimisation problem with the two objectives being the minimisation of the number of sites involved in processing the distributed query and the maximisation of the concentration of relations in these involved sites. This problem has been solved using the set-based comprehensive learning parallel particle swarm optimisation (S-CLPPSO). Experimental results show that the S-CLPPSO-based distributed query plan generation algorithm is able to generate Top-K query plans that would result in efficient processing of a distributed query.
Keywords: distributed databases; query processing; query plans; multiobjective optimisation; parallel PSO; particle swarm optimisation; PPSO; distributed queries.
International Journal of Collaborative Intelligence, 2015 Vol.1 No.2, pp.85 - 114
Received: 23 Dec 2014
Accepted: 20 Jan 2015
Published online: 17 Aug 2015 *