Title :
A new distributed evolutionary computation technique for solving large number of equations
Author :
Jahan, Moslema ; Hashem, M.M.A. ; Shahriar, Gazi Abdullah
Author_Institution :
Dept. of Comput. Sci. & Eng., Khulna Univ. of Eng. & Technol., Khulna
Abstract :
Evolutionary algorithm is more effective to gain optimal solution to solve equations than traditional methods. It also provides quick solution for solving large number of equations having huge parameters with less expense in distributed manner. This paper presents a new distributed evolutionary computation technique, which decomposes decision vectors into smaller components and achieves optimal solution in short time. In this technique, A Jacobi-based Time Variant Adaptive (JBTVA) Hybrid Evolutionary Algorithm is distributed. Moreover, a new selection method named Best All Selection (BAS) is introduced for selecting best individuals. Experimental results show that optimal solution is achieved for different kinds of problems having huge parameters and considerably speedup is formed in proposed distributed system.
Keywords :
distributed processing; evolutionary computation; Jacobi-based time variant adaptive; best all selection; distributed evolutionary computation; distributed system; hybrid evolutionary algorithm; optimal solution; Computational modeling; Computer architecture; Computer science; Distributed computing; Equations; Evolutionary computation; Genetic mutations; Information technology; Jacobian matrices; Master-slave; Evolutionary Algorithms; Hybrid Algorithm; Linear Equations; Master-Slave architecture; Speedup;
Conference_Titel :
Computer and Information Technology, 2008. ICCIT 2008. 11th International Conference on
Conference_Location :
Khulna
Print_ISBN :
978-1-4244-2135-0
Electronic_ISBN :
978-1-4244-2136-7
DOI :
10.1109/ICCITECHN.2008.4802989