Periodic Block Tridiagonl System
Hi, Can you tell me where I can get the knowledge of the solution of Periodic Block Tridiagonl System of equations? Thank you!
A.M.Yang 02/27/01 |
Re: Periodic Block Tridiagonl System
|
Re: Periodic Block Tridiagonl System
Try this book:
Anderson, Tannehill, Pletcher COMPUTATIONAL FLUID MECHANICS AND HEAT TRANSFER, Hemisphere Publishing Corporation(1984),p.551 Regards |
Re: Periodic Block Tridiagonl System
Dear Mr.Barbaro,M:
Thank you for your advice. The Book of Anderson is at my hand, but the program seems erroneous. Regards A.M.Yang 02/27/2001 |
Thank You!
Dear Mr. Sylvain:
Thank you for your advice! Regards A.M.Yang 02/27/2001 |
Re: Periodic Block Tridiagonl System
I had the same problem a few years ago. It was not in my copy of Numerical Recipes (1986) but may be in a newer version (?). An approach is as follows (I never found a reference):
Write down the matrix problem. Pull out the last equation for Xn to leave a tridiagonal system. Split the unknown x into the sum of two parts: x = f + Xn*g. Solve two TDMA problems for f and g (g gets the Xn*A terms to divide out Xn and the f the rest). Use Xn equation to get Xn. Evaluate x from f and g. |
Re: Periodic Block Tridiagonl System
Dear Andy:
Thank you for your advice! Regards A.M.Yang |
Re: Periodic Block Tridiagonl System
Dear Andy:
Can you give me the process in detail? Thank you very much! Regards A.M.Yang (aimingyang@263.net) |
Re: Periodic Block Tridiagonl System
A reference has been pointed out to me. Fortran code is contained in an appendix of Volume 1 of Hirsch's book. My original source was a proprietary routine from the 1970s. They have coding similarities which suggests it is in a numerical library somewhere or, perhaps, it is just one of those routines that has been passed among Ph.D. students since the dawn of time. Digging about on netlib is likely to be fruitful. However, I had missed the word "block" on the first reading of your question so it may not be want you want anyway.
|
Re: Periodic Block Tridiagonl System
Dear Andy:
Thank you very much ! Regards A.M.Yang |
All times are GMT -4. The time now is 03:39. |