Send mail to:             for the digests or bakeoff
            for comments or help
 Current editor:  Craig Douglas       
Anonymous ftp repository: (

World Wide Web: or

Today's editor:  Craig Douglas (

Volume 6, Number 11 (approximately November 30, 1996)

Today's topics:

     MGNet Web Page Reorganization
     Request for Information
     3 new papers (Yavneh)
     New Paper (Shapira)
     8th Copper Mountain Conference on Multigrid Methods Reminder
     Additions to mgnet.bib (Tai)
     Additions to mgnet.bib (Brenner)
     Another Faculty Position at the University of Kentucky


Date: Sat, 30 Nov 1996 07:58:41 -0500 (EST)
From: Craig Douglas 
Subject: MGNet Web Page Reorganization

Over the past month, many of the web pages for MGNet have undergone an
extensive set of updates.  This will continue over the next 2 months as I find

A professional art designer offered a number of specific suggestions.  It is
hopefully now much easier to find information and visually more appealing.  It
is very easy to tell which parts have been changed and which parts have yet to
be updated.  The principal pages have all undergone revision.  Suggestions are

One of the changes that will go into effect on December 3 is that there will
be a "link of the week" to a web site that I think is of interest to the
readers of this digest.  If you want to see what has been picked, you have to
visit MGNet's main web page and click on the link.  It will usually change on
Monday thereafter.

If you think that your personal or group web page is noteworthy, please send
me the URL by e-mail.

Chris Johnson of the University of Utah was the first to report a problems
with the new pages.  It turned out that Netscape's browser handles JavaScript
a little differently than is published.  All of the effected pages have been
fixed to get around this anomoly.  My pages caused a core dump on his SGI.
Please contact me if you have a problem.

If you have not already done so, this is a good time to upgrade to version 3+
of either Netscape's or Microsoft's browsers.  Also, the pages look better on
color displays if you go into the general preferences/appearances options menu
and turn off automatic underlining of links.


Date: Sat, 30 Nov 1996 08:10:12 -0500 (EST)
From: Craig Douglas 
Subject: Request for Information

Essentially all of the preprints and virtual proceedings' papers on MGNet are
stored as gzipped PostScript files.  For anyone using a UNIX machine to view
these papers with netscape, it is easy to set up a .mailcap file to
automatically gunzip a paper and start a viewer like ghostview.

A number of people have recently asked how to do this when using a Windows 95
machine or a Mac.  I will admit to virtually never using either (many of you
know that I use Linux on my PC's).

If you have this problem solved, could you please e-mail me your solution?  It
does not really matter if I get multiple solutions, as long as I can post ones
that work for Windows and Macs.

Thanks in advance... Craig


Date: Wed, 27 Nov 1996 11:16:18 +0200
Subject: 3 new papers (Yavneh)

I just put three papers in the "incoming" directory, two in 
subdirectory Yavneh, and one in Yavneh-Venner-Brandt. The papers are
in gzipped postscript and the abstracts in ascii. Please let me know
if anything else is needed before these can be added to my existing

Irad Yavneh.

                 Coarse-Grid Correction for Nonelliptic and 
                        Singular Perturbation Problems

                                 Irad Yavneh 
                       Department of Computer Science 
                  Technion---Israel Institute of Technology
                             Haifa 32000, Israel
                                 October 1996

The suitability of coarse-grid correction operators for multigrid solution of
discretized nonelliptic partial differential equations is studied.  It is
found that the requirements for the orders of intergrid transfers and the
order of approximation of the coarse-grid operators differ in an important way
from the well-known ``rules'' pertaining to elliptic operators.  Both Galerkin
coarsening and coarsening schemes based on the differential operator are
examined, and new coarse-grid correction strategies are suggested in light of
this analysis.

    Editor's Note: in mgnet/papers/Yavneh/

            Multigrid Smoothing for Symmetric Nine-Point Stencils

                        Irad Yavneh and Elena Olvovsky
                       Department of Computer Science 
                  Technion---Israel Institute of Technology
                             Haifa 32000, Israel
                                November 1996

Analytical formulae are obtained for the smoothing factors yielded by damped
Jacobi relaxation and by red-black relaxation applied to symmetric nine-point
stencil discretizations of elliptic partial differential operators in 2D.  The
results include point and line relaxation and full and partial coarsening.
Several unusual results are implied by the formulae.  Numerical results of
multigrid cycles match the analytical predictions well.

    Editor's Note: in mgnet/papers/Yavneh/

               Fast Multigrid Solution of the Advection Problem
                         with Closed Characteristics

                                  I. Yavneh
           Technion---Israel Institute of Technology Haifa, Israel
                                 C.H. Venner
               University of Twente, Enschede, The Netherlands
                                  A. Brandt
                Weizmann Institute of Science, Rehovot, Israel
                                September 1995

The numerical solution of the advection-diffusion problem in the inviscid
limit with closed characteristics is studied as a prelude to an efficient high
Reynolds-number flow solver.  It is demonstrated by a heuristic analysis and
numerical calculations that using upstream discretization with downstream
relaxation-ordering in a multigrid cycle with appropriate residual weighting
leads to an efficient solution process.  Upstream finite-difference
approximations to the advection operator are derived whose truncation terms
approximate ``physical'' (Laplacian) viscosity, thus avoiding spurious
solutions to the homogeneous problem when the artificial diffusivity dominates
the physical viscosity.

    Editor's Note: in mgnet/papers/Yavneh-Venner-Brandt/


Date: Mon, 2 Dec 1996 15:28:56 -0700 (MST)
From: Yair Shapira 
Subject: New Paper (Shapira)

                     Multigrid for Locally Refined Meshes

                                 Yair Shapira

          Computer Science Department, Technion, Haifa 32000, Israel
                    Now at Los Alamos National Laboratory
                        MS B-256, Los Alamos NM 87545


A two-level method for the solution of finite element schemes on locally
refined meshes is introduced.  The upper bound on the condition number
implies, in some cases, mesh-independent convergence, as is verified
numerically for a diffusion problem with discontinuous coefficients.  The
discontinuity curves are not necessarily aligned with the coarse mesh; indeed,
numerical applications with ten levels of local refinement yield a fast
convergence for the corresponding ten-level multigrid V-cycle, even when the
discontinuities are invisible on most of the coarse meshes.

    Editor's Note: in mgnet/papers/Shapira/


Date: Fri, 29 Nov 1996 20:41:25 -0700 (MST)
From: Steve McCormick 
Subject: 8th Copper Mountain Conference on Multigrid Methods Reminder

Copper Mountain, Colorado, USA
APRIL 6-11, 1997

Travel and lodging assistance will be awarded to student authors judged to
have the best research papers.  Papers containing original research due mainly
to the student must be received before Dec.  15, 1995.  They must be singly
authored and may be no more than 10 pages in length.  Sending only an abstract
is unacceptable.  Authors will be expected to present their paper in one of
the regular conference sessions.  Student papers should be submitted as ps
files by e-mail to  If this is not possible, a hard
copy of the paper may be submitted by regular post to:
    Applied Math
    Campus Box 526
    University of Colorado
    Boulder, CO 80309-0526

Author Abstracts                      January 15, 1997
Papers for NASA Proceedings           March 15, 1997
Guaranteed Availibility of Lodging    February 6, 1997
Early Registration                    February 6, 1997

For information such as registration, author instructions, and lodging
arrangements (including a lodging bulletin board), please access:  You may also contact
us by email at

Steve McCormick: Appl. Math, C.B. 526, U. of CO, Boulder, CO 80309-0526
        -4066 fax


Date: Fri, 1 Nov 96 15:13:03 MET
Subject: Additions to mgnet.bib (Tai)

Could you please add the following of my works into you mgnet.bib file?

Sincerely Tai.

   author = "Xue-Cheng Tai and P. Neittaanmaki",
    title = "A parallel finite element splitting-up methods for parabolic problem",
  journal = "Numerical Methods for Partial Differential Equation",
     year = "1991",
   volume = "7",
    pages = "209--225",
   author = "T. Lu and P. Neittaanmaki and X-C Tai",
    title = "A parallel splitting up method and its application to Navier-Stoke equations",
  journal = "Applied Mathematics Letters",
     year = "1991",
   volume = "4",
    pages = "25--29",
   author = "T. Lu and P. Neittaanmaki and X-C Tai",
    title = "A parallel splitting up method for partial differential equations and its application to Navier-Stokes equations",
  journal = "RAIRO Math. Model. and Numer. Anal.",
     year = "1992",
   volume = "26",
    pages = "673--708",
   author = "Xue-Cheng Tai and Pekka Neittaanmaki",
    title = "Error estimates for numerical identification of distributed parameters",
  journal = "J. Comp. Math.",
     year = "1992",
   volume = "supplementary issue",
    pages = "66-78",
   author = "Xue-Cheng Tai",
    title = "Global extrapolation with a parallel splitting method",
  journal = "Numerical Algorithms",
     year = "1992",
   volume = "3",
   pages  ="427--440",
   author = "X.-C. Tai and T. Karkainen",
    title = "Identification of a nonlinear parameter in a parabolic equation from a linear equation",
  journal = "Comp. Appl. Mat.",
     year = "1995",
   volume = "14",
    pages = "157--184",
   author = "Xue-Cheng Tai",
    title = "Parallel function and space decomposition methods -- {Part I}. Function decomposition",
  journal = "Beijing Mathematics",
     year = "1995",
   volume = "1, part 2",
    pages = "104--134",
   author = "Xue-Cheng Tai",
    title = "Parallel function and space decomposition methods -- {Part II}. Space decomposition",
  journal = "Beijing Mathematics",
     year = "1995",
   volume = "1, part 2",
    pages = "135--152",
   author = "K. Kunisch and X.-C. Tai",
    title = "Sequential and parallel splitting methods for bilinear control problems in Hilbert spaces",
  journal = "SIAM J. Numer. Anal.",
     year = "1997",
   author = "Xue-Cheng Tai and Pekka Neittaanmaki",
    title = "A {FE-splitting-up} method and its application to distributed parameter identification of parabolic type",
 booktitle= "Approximation, Optimisation and  Computing. Theory and Applications",
   editor = "A. G. Law and C. L. Wang",
     year = "1990",
    pages = "185--188",
publisher = "Elsevier",
   author = "Xue-Cheng Tai and Pekka Neittaanmaki",
    title = "A linear approach for the nonlinear distributed parameter identification",
booktitle = "Numerical methods for free boundary problems",
     year = "1991",
   volume = "99",
   editor = "P. Neittaanmaki",
    pages = "401-411",
publisher = "Birkhauser Verlag Basel",
   series = "International series of numerical mathematics",
   author = "Xue-Cheng Tai and Pekka Neittaanmaki",
    title = "On the numerical solution of the distributed parameter parameter identification problem",
booktitle = "Control and estimation of distributed parameter systems",
     year = "1991",
   volume = "100",
    pages = "317--329",
publisher = "Birkhauser Verlag Basel",
   series = "International series of numerical mathematics",
   author = "Xue-Cheng Tai",
    title = "Domain decomposition for linear and nonlinear elliptic problems via function or space decomposition",
booktitle = "Domain decomposition methods in scientific and engineering computing",
     year = "1995",
   editor = "D. Keyes and J. Xu",
    pages = "355--360",
publisher = "American Mathematical Society",
   series = "Proc. of the 7th international conference on domain decomposition, Penn. State University, 1993",
   author = "Xue-Cheng Tai",
    title = "Parallel function and space decomposition methods",
booktitle = "The Finite element Method: Fifty years of the Courant Element",
     year = "1994",
   volume = "164",
   editor = "P. Neittaanmaki",
    pages = "421--432",
publisher = "Marcel Dekker",
   series = "Lecture notes in pure and applied mathematics",

Dr Xue-Cheng Tai                           Tel. 47 55 58 4868 (office), 
Bergen University,                         Tel. 47 55 34 7356 (home),
Mathematics Department,                    Fax. 47 55 58 4885,
Alleg. 55, 5007, Bergen, Norway.           www:

    Editor's Note: Now in mgnet/bib/mgnet.bib.


Date: Mon, 2 Dec 1996 14:20:46 -0500
From: Susanne Brenner 
Subject: Additions to mgnet.bib (Brenner)

  author =      "P. Grisvard", 
  title =       "Singularities in Boundary Value Problems",
  series =      "Research Notes in Applied Mathematics",
  volume =      "22",
  publisher =   "Masson",
  address =     "Paris",
  year =        "1992",
  author =      "Eugene G. D'yakonov", 
  title =       "Optimization in Solving Elliptic Problems",
  series =      "",
  volume =      "",
  publisher =   "CRC Press, Inc.",
  address =     "Boca Raton",
  year =        "1996",

    Editor's Note: Will be in mgnet/bib/mgnet.bib shortly.


Date: Sat, 30 Nov 1996 10:11:12 -0500 (EST)
From: Craig Douglas 
Subject: Another Faculty Position at the University of Kentucky

While I was visiting Lexington recently, I discovered that the computer
science department is trying to hire a faculty member in scientific computing.
This is in addition to the math department position that was listed here last
month.  If you are interested in the details, look at

for the real details.  (Note: I am just passing this information along.)

If a reader of this mailing list applies for either the math or the computer
science positions, would you drop me a one line note, please?  Thanks.


End of MGNet Digest