Computational Technology Resources - CCP

B.H.V. Topping and A. Bahreininejad

Department of Mechanical and Chemical Engineering, Heriot-Watt University, Edinburgh, UK

doi:10.4203/ccp.34.3.2

B.H.V. Topping, A. Bahreininejad, "Subdomain Generation using Parallel Q-State Potts Neural Networks", in B.H.V. Topping, (Editor), "Developments in Neural Networks and Evolutionary Computing for Civil and Structural Engineering", Civil-Comp Press, Edinburgh, UK, pp 65-78, 1995. doi:10.4203/ccp.34.3.2

Abstract

The implementation of parallel unsupervised recurrent neural networks to the partitioning of unstructured adaptive meshes for domain decomposition of finite element meshes for explicit time-stepping dynamic analysis is examined. The use of the Mean Field Annealing (MFA) technique and the multi-state Potts model for finding approximate solutions to the partitioning of the finite element meshes is investigated. The method of mapping the subdomain generation problem, both in predictive and non-predictive modes, onto the Potts neural network and the parallelisation of the annealing procedure are presented. The solution qualities are compared with the Subdomain Generation Method based on the Genetic Algorithm (GASGM) and the Greedy Algorithm. All computational studies were carried out using T800 transputers.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £55 +P&P)

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Proceedings ISSN 1759-3433 CCP: 34 DEVELOPMENTS IN NEURAL NETWORKS AND EVOLUTIONARY COMPUTING FOR CIVIL AND STRUCTURAL ENGINEERING Edited by: B.H.V. Topping Paper III.2 Subdomain Generation using Parallel Q-State Potts Neural Networks B.H.V. Topping and A. Bahreininejad Department of Mechanical and Chemical Engineering, Heriot-Watt University, Edinburgh, UK doi:10.4203/ccp.34.3.2 purchase the full-text of this paper Full Bibliographic Reference for this paper B.H.V. Topping, A. Bahreininejad, "Subdomain Generation using Parallel Q-State Potts Neural Networks", in B.H.V. Topping, (Editor), "Developments in Neural Networks and Evolutionary Computing for Civil and Structural Engineering", Civil-Comp Press, Edinburgh, UK, pp 65-78, 1995. doi:10.4203/ccp.34.3.2 Abstract The implementation of parallel unsupervised recurrent neural networks to the partitioning of unstructured adaptive meshes for domain decomposition of finite element meshes for explicit time-stepping dynamic analysis is examined. The use of the Mean Field Annealing (MFA) technique and the multi-state Potts model for finding approximate solutions to the partitioning of the finite element meshes is investigated. The method of mapping the subdomain generation problem, both in predictive and non-predictive modes, onto the Potts neural network and the parallelisation of the annealing procedure are presented. The solution qualities are compared with the Subdomain Generation Method based on the Genetic Algorithm (GASGM) and the Greedy Algorithm. All computational studies were carried out using T800 transputers. purchase the full-text of this paper (price £20) go to the previous paper go to the next paper return to the table of contents return to the book description purchase this book (price £55 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions