**Today's Topics:**

- Oops, Two Issues Numbered 5
- SIAM SIAG Supercomputing Newsletter
- Polynomial Zeros in C
- Numerical Analysis Position at Johns Hopkins University
- Special Issue of Mathematical Programming
- Congratulations Alan George
- New Address for Margaret Wright
- MILP/MINLP Software

From: Cleve Moler <moler@na-net.stanford.edu>

Date: Sat Feb 13 07:52:23 PST 1988

We've recently switched the NA Net source machine at Stanford

from "score", an old DEC 10 or 20 running TOPS, to "na-net",

a Sun running UNIX. (Thanks to Mark Kent for making the

switch.) In the process, I forget to increment the issue

number, so there have been two issues numbered 5. I'm calling

this issue number 7 in case anybody else is saving back issues

and wants room to call one of those issue number 6.

--Cleve

------------------------

From: Aaron J. Owens <OWENS%ESVAX%dupont.com@RELAY.CS.NET>

Date: Tue, 9 Feb 88 14:20 EDT

As co-editors of the newsletter for the SIAM Special Interest

Activity Group on Supercomputing, we welcome contributions (eg.,

announcements, meeting reports, articles, book reviews) for the

newsletter. Items of general interest to be submitted to SIAM News

are also solicited.

Electronic, hardcopy, or telephone contributions will be gladly

accepted. Please address electronic mail on CSNET to

owens%esvax@dupont.com

and analog mail to

A. J. Owens

CR&DD, E320

Du Pont Co.

Wilmington, DE 19898

-- Dave Filkin -- Telephone (302) 695-3970

-- Aaron Owens -- Telephone (302) 695-1762

------------------------

From: Tony Mestoussis <mcvax!prlb2!lln-cs!etbel!root@uunet.uu.net>

Date: 8 Feb 88 15:22:01 GMT

Organization: Electronic and Telecommunication Bell ,Namur ,BELGIUM

I am looking for a good algorithm for computing all the

rational zeros of a polynomial with real coefficients. The

polynomial power, the value of each coefficient and the

absolute value of the acceptable error are given by the user.

Initial information about the zeros is not available.

The highest power is 127 and the maximum error is 10exp(-15).

The program will be written in C and it is assumed that the

target machines have a 64-bit double float including 10+1 bits

for the exponant and 52+1 bits for the mantissa and that

correct rounding is used.

If you can help me please let me know via any mean you see fit.

Thanks in advance.

Tony Mestoussis

Rue Pierre du Diable, 34

B5100 NAMUR (Jambes)

BELGIUM

Private Tel. : International : 00 32 81 302922

National : (081) 302922

Office Tel. : International : 00 32 81 304001

National : (081) 304001

uucp : ...!mcvax!prlb2!etbel!tony

------------------------

From: Roger Horn <MSC_HRAH%JHUVMS.BITNET@forsythe.stanford.edu>

Date: Wed, 10 Feb 88 08:33 EST

NUMERICAL ANALYSIS POSITION

DEPARTMENT OF MATHEMATICAL SCIENCES

THE JOHNS HOPKINS UNIVERSITY

BALTIMORE, MARYLAND 21218

The Department of Mathematical Sciences at Johns Hopkins has an open tenure-

track junior position in numerical analysis/optimization. Recruiting is now

underway, so anyone interested should communicate with us right away. Direct

inquiries from prospective candidates are welcome, as are referrals from third

parties. For rapid response to informal inquiries contact Roger Horn at

NA.HORN@SCORE.STANFORD.EDU or directly through BITNET at MSC_HRAH@JHUVMS (note

that the underscore is essential). To apply formally, write to Professor

Robert Serfling, Chairman.

------------------------

From: Andrew Roger Conn <arconn%water.waterloo.edu@RELAY.CS.NET>

Date: Thu, 11 Feb 88 22:14:22 EST

MATHEMATICAL PROGRAMMING

CALL FOR PAPERS

A special issue of Mathematical Programming Series B devoted

to large scale problems is being prepared under the joint editorship of

A.R.Conn, N.I.M.Gould and Ph.L.Toint..

The emphasis is intended to be towards large scale nonlinear programming

rather than linear programming. However, articles that emphasize applications,

algorithms and/or theory that is not primarily linear programming oriented

could be suitable and are solicited.

Mathematical Programming Series B is now on a similar schedule to Series A

with similar standards for publication and the same circulation.

The advantage of a special issue, besides collecting articles on a coherent

theme is that the refereeing process can often be expedited.

If you have a suitable paper for such an issue you are invited to submit

it to one of the editors

A.R.Conn

Department of Combinatorics and Optimization

Faculty of Mathematics

University of Waterloo

Waterloo

Ontario

Canada N2L 3G1

Tel:519-888-4458 E-mail: arconn@water.bitnet

arconn@water.waterloo.edu

N.I.M.Gould

Computer Science and Systems Division

A.E.R.E. Harwell

Oxfordshire OX11 ORA

U.K.

Tel:235-24141 extn3357 E-mail: nimgould@water.bitnet

nimgould@water.waterloo.edu

Ph.L.Toint

Department of Mathematics

Facultes Universitaires Notre Dame de la Paix

Rue de Bruxelles, 61

B-5000

Namur

Belgium

Tel:32-81-229061 extn 2592 E-mail: phtoint@bnandp10.bitnet

The deadline for submission is August 1 1988 for the issue that is to be

published in 1989.

Late submissions, an excess of submissions, etc., would automatically be

considered for Mathematical Programming Series A, unless the author wishes

otherwise.

------------------------

From: Gene Golub <golub@golubsun.cs.umd.edu>

Date: Mon, 8 Feb 88 13:34:54 EST

Alan George will be leaving UTK/ORNL to assume the position of Provst

and Vice-President at the University of Waterloo effective July 1,

1988.

--Gene

------------------------

From: Margaret Wright <OR.WRIGHT@Sierra.Stanford.EDU>

Date: Fri 12 Feb 88 11:21:34-PST

As of February 15, 1988, my address, phone and

e-mail address will be:

Margaret H. Wright

AT&T Bell Laboratories

Room 2C-462

600 Mountain Avenue

Murray Hill, New Jersey 07974

Telephone: 201-582-3498

E-mail: mhw@research.att.com

or

na.mwright@score.stanford.edu (the same as now)

------------------------

From: Himadri Das <uvaarpa!hudson!gibbs!dh4v@umd5.umd.edu>

Date: 12 Feb 88 18:28:57 GMT

Organization: University of Virginia, Charlottesville

I am looking for software packages to numerically solve MILP (Mixed Integer

Linear Programming) and MINLP (Mixed Integer Non Linear Programming)

optimization problems.

I would appreciate it if you could tell me what are the available

packages and how I could lay my hands on them.

Please send e mail to HD7Q@VIRGINIA.BITNET

Himadri Das

Dept of Chemical Engineering

Thornton Hall

University of Virginia

Charlottesville, VA 22901

Phone: (804)-924-6287

------------------------------

End of NA Digest

**************************

-------