A Distributed Genetic Algorithm Solution to the Boolean Satisfiability Problem

HASAN, Mohammad, Amavasai, Bala P and Travis, J R (2005) A Distributed Genetic Algorithm Solution to the Boolean Satisfiability Problem. In: International Symposium on Innovations in Intelligent Systems and Applications.

Preview

Text GA_BSAT_inista_05.pdf - AUTHOR'S ACCEPTED Version (default) Available under License All Rights Reserved. Download (193kB) | Preview

Abstract or description

This paper attempts to improve the solution of the NP complete Boolean Satisfiability (BSAT) problem by partitioning the task into three sub-tasks and distributing them over an experimental 3-node Distributed Computing System (DCS). A genetic algorithm (GA) has been used to consider multiple feasible solutions. The GA based algorithm is applied to the standard BSAT benchmarks on a single computer and on DCS configuration using non-optimised and optimised executables. The task is coarsely partitioned and distributed over the DCS using the Simple Object Access Protocol (SOAP) technology. The results reveal that the DCS enabled solution exhibits better performance than a single computer configuration for non-optimised GA code. However, no clear correlation could be identified between the single computer and the DCS for the optimised version of the GA search. The main contribution of this investigation is the design of a GA based solution to the BSAT problem for DCS.

Item Type:	Conference or Workshop Item (Paper)
Additional Information:	© 2005 IEEE
Faculty:	School of Computing and Digital Technologies > Computing
Event Title:	International Symposium on Innovations in Intelligent Systems and Applications
Depositing User:	Mohammad HASAN
Date Deposited:	12 Mar 2019 15:10
Last Modified:	24 Feb 2023 13:55
URI:	https://eprints.staffs.ac.uk/id/eprint/5453

Actions (login required)

View Item