Correcting Interface Errors Arising from Segmentation in a Parallel Iterative Algorithm
Several parallel iterative strategies based on domain segmentation were previously developed by the authors. While each is effective, the error distribution after every global iteration is not uniform within each segment, the error being greater at the extremities. A correction to the solution is developed at the interfaces between the disjoint segments which accelerates convergence.
Unable to display preview. Download preview PDF.
- Barry, J.M., Pollard, J.P. and Wachspress E.L. (1988a) - Some parallel iterative schemes for large systems of linear equations arising in three dimensional modelling - to appear.Google Scholar
- Barry, J.M., Pollard, J.P. and Wachspress, E.L. (1988b) - A method of parallel iteration - to appear.Google Scholar
- Barry, J.M. and Wachspress, E.L. (1988) - A method of parallel iteration based on polynomial correction in Noye J. and Fletcher C., Computational techniques and applications CTAC-87, North Holland, pp. 93–98.Google Scholar
- Mahoney, M.A. and Wachspress, E.L. (1985) - Numerical solution of 3D neutron group diffusion equations on the CYBER-205 computer, in Proceedings of international meeting on advances in nuclear engineering computational methods, American Nuclear Society (1), pp. 66–71.Google Scholar