From thouis at graphics.csail.mit.edu Mon Dec 1 08:12:38 2003
From: thouis at graphics.csail.mit.edu (Ray Jones)
Date: Mon Jan 9 13:41:12 2006
Subject: Relative Distance Cartogram algorithm question
InReplyTo: <20031201025138.17278.qmail@web60310.mail.yahoo.com>
References: <20031201025138.17278.qmail@web60310.mail.yahoo.com>
MessageID:
Boris Dev writes:
> If we have a distance matix can we put points on a 2
> diminsional x, y coordinate grid so that they are
> postioned relative to one another according to the
> distance matrix elements.
Search for "multidimensional scaling". Or follow the links and
references from this site and the paper it refers to:
http://isomap.stanford.edu/
Ray Jones

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From michael.aupetit at cea.fr Mon Dec 1 13:25:57 2003
From: michael.aupetit at cea.fr (aupetit)
Date: Mon Jan 9 13:41:12 2006
Subject: Relative Distance Cartogram algorithm question
References: <20031201025138.17278.qmail@web60310.mail.yahoo.com>
MessageID: <200312011226.NAA29388@tupai.bruyeres.cea.fr>
Try these methods or keywords
MultiDimensional Scaling (MDS)
or Sammon's NonLinear Mapping
and efficient variations of it such as
Curvilinear Components Analysis (Demartines, Herault)
Hope this helps
Michael
Boris Dev wrote:
> I hope this is a comp geom problem:
>
> If we have a distance matix can we put points on a 2
> diminsional x, y coordinate grid so that they are
> postioned relative to one another according to the
> distance matrix elements.
>
>  what if distance was defined in some noneuclidian
> terms based on say correlation coeffients between
> composite units of an aggregate (say as a function of
> USA states'
> comovements). In this case all restrictions based on
> data might not be met with 2 dim coordinate plane. So
> will
> 3dimensions suffice?
>
> Is there an algorithm out there?
>
> Any advice?
>
> Ultimately, I want to make a graph/cartogram based on
> different relative measures of distance.
>
> Thanks much for all your time.
> borisdev@yahoo.com
>
> 
> The compgeom mailing lists: see
> http://netlib.belllabs.com/netlib/compgeom/readme.html
> or send mail to compgeomrequest@research.belllabs.com with the line:
> send readme
> Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
 next part 
A nontext attachment was scrubbed...
Name: aupetit.vcf
Type: text/xvcard
Size: 272 bytes
Desc: Card for aupetit
Url : http://compgeom.poly.edu/pipermail/compgeomannounce/attachments/20031201/99338414/aupetit.vcf
From large at cs.duke.edu Tue Dec 2 20:47:26 2003
From: large at cs.duke.edu (Lars Arge)
Date: Mon Jan 9 13:41:12 2006
Subject: ALENEX'04 accepted papers
MessageID: <3FCD40AE.3090600@cs.duke.edu>
Papers accepted for presentation at the
6th Workshop on Algorithm Engineering and Experiments (ALENEX'04).
January 10, 2004, Astor Crown Plaza Hotel, New Orleans, Louisiana.
http://www.siam.org/meetings/alenex04.

Engineering a CacheOblivious Sorting Algorithm
Gerth Brodal, Rolf Fagerberg and Kristoffer Vinther
The Robustness of the SumofSquares Algorithm for Bin Packing
M. Bender, B. Bradley, G. Jagannathan and K. Pillaipakkamnatt
Practical Aspects of Compressed Suffix Arrays and FMindex in Searching
DNA Sequences
WingKai Hon, TakWah Lam, WingKin Sung, WaiLeuk Tse, ChiKwong Wong
and SiuMing Yiu
Faster placement of hydrogens in protein structures by dynamic programming
Andrew LeaverFay, Yuanxin Liu and Jack Snoeyink
An Experimental Analysis of a Compact Graph Representation
Dan Blandford, Guy Blelloch and Ian Kash
Kernelization Algorithms for the Vertex Cover Problem: Theory and
Experiments
Faisal N. AbuKhzam, Rebecca L. Collins, Michael R. Fellows and Michael
A. Langston
Safe Separators for Treewidth
Hans L. Bodlaender and Arie M.C.A. Koster
Efficient Implementation of a Hotlink Assignment Algorithm for Web Sites
Artur Alves Pessoa, Eduardo Sany Laber and Cr?ston de Souza
Experimental Comparison of Shortest Path Approaches for Timetable
Information
Evangelia Pyrga, Frank Schulz, Dorothea Wagner and Christos Zaroliagis
Reachbased Routing: A New Approach to Shortest Path Algorithms
Optimized for Road Networks
Ron Gutman
Lazy Algorithms for Dynamic Closest Pair with Arbitrary Distance Measures
Jean Cardinal and David Eppstein
Approximating the Visible Region of a Point on a Terrain
Boaz BenMoshe, Paz Carmi and Matthew J. Katz
A computation framework for handling motion
Leo Guibas, Menelaos Karavelas and Daniel Russel
Engineering a Sorted List Data Structure for 32 Bit Keys
Roman Dementiev, Lutz Kettner, Jens Mehnert and Peter Sanders

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From h.j.gudmundsson at tue.nl Fri Dec 5 15:16:29 2003
From: h.j.gudmundsson at tue.nl (Joachim Gudmundsson)
Date: Mon Jan 9 13:41:12 2006
Subject: CG models of computation?
MessageID: <000301c3bb3a$60ea4910$84449b83@campus.tue.nl>
Dear Colleagues, I have a question about which models
of computation that are accepted in the computational
geometry community.
In many of the papers I have coauthored we have
strived to use the algebraic decision tree model of
computation (sometimes extended with indirect
addressing), but sometimes also using the real RAM.
Recently, I read an excellent paper where the authors
add the power of the floor function to the standard
model of computation. The running time of their
algorithm is almost linear and beats the lower bound
in the comparison model.
Somewhat puzzled and skeptical I asked my colleagues
about their opinion on this topic, only to find out that
this model was, according to my colleagues, unofficially
accepted. One of my colleagues even claimed that it was
so widely accepted that it is not even necessary to
explicitly point out in a paper if the floor function
is used or not. Amazed about this information I went
back to my office and looked at a paper I'm currently
working on. Adding the floor function would trivially
improve the running time of the algorithm from
O((m+n) log n) to O(m+n log n), maybe even more. My
question is, what is "unofficially" allowed? One could
claim that the use of floor function should be accepted
since it is used in practice and very fast. How about
bit manipulation? But then we should measure the
complexity in the number of input bits, right?
Anyway, I would be very happy if anyone could help remove
my obvious ignorance on this topic.
Sincerely,
Joachim Gudmundsson

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From jeffe at cs.uiuc.edu Sun Dec 7 04:16:10 2003
From: jeffe at cs.uiuc.edu (Jeff Erickson)
Date: Mon Jan 9 13:41:12 2006
Subject: CG models of computation?
InReplyTo: <000301c3bb3a$60ea4910$84449b83@campus.tue.nl>
References: <000301c3bb3a$60ea4910$84449b83@campus.tue.nl>
MessageID: <20031207101610.GI9674@granmapa.cs.uiuc.edu>
Joachim Gudmundsson wrote:
 Recently, I read an excellent paper where the authors add the power
 of the floor function to the standard model of computation. The
 running time of their algorithm is almost linear and beats the lower
 bound in the comparison model.

 Somewhat puzzled and skeptical I asked my colleagues about their
 opinion on this topic, only to find out that this model was,
 according to my colleagues, unofficially accepted.

 My question is, what is "unofficially" allowed? One could claim
 that the use of floor function should be accepted since it is used
 in practice and very fast. How about bit manipulation? But then we
 should measure the complexity in the number of input bits, right?
This is actually a pretty big can of worms!
The most common model of geometric computation seems to be the
unitcost real RAM, but it is fairly common to add the floor function
to permit bucketing or hashing, in many cases beating lower bounds in
the algebraic decision tree model. The two classical examples are
Gonzalez's algorithm for MAX GAP [gasrp75] and Rabin's randomized
algorithm for closest pairs [rpa76].
However, despite its general acceptance, the unitcost real RAM with
the floor function is NOT a reasonable model of computation, because
it allows any problem in PSPACE or #P to be solved in polynomial time!
In 1979, Sch\"onhage [spram79] described an algorithm to solve the
PSPACEcomplete problem QBFdeciding if a given arbitrarily
quantified boolean formula is true or falseusing a polynomial
number of integer arithmetic operations: z=x+y, z=xy, z=xy, and
z=floor(x/y). The trick is to encode the entire formula into a single
integer and then use arithmetic to process different parts of the
encoded forumla in parallel. His algorithm just removes the
quantifiers, by replacing each ExF(x) with F(0)vF(1) and each AxF(x)
with F(0)^F(1), and then simplifies the resulting quantifierfree
formula to either 0 or 1. Hartmanis and Simon [hspmram74] did the
same thing in 1974, only using bitwise Boolean operations instead of
integer division. A few years later, Bertoni et al. [bmsscram85]
generalized the same approach to the #Pcomplete problem #SAT: How
many satisfying assignments does this boolean formula have? Peter van
Emde Boas has a great discussion of "the unreasonable power of integer
multiplication" in his survey of models of computation [emms90].
Partly as a result of the HartmanisSimon result, there are now two
essentially standard integer RAM models:
(1) Logarithmiccost (or "bitlevel") RAM: Each memory location can
store an arbitrary integer. The cost of each arithmetic operation
is proportional to the total number of BITS involved.
(2) Wordlevel RAM: Each memory location can store a single word
consisting of Theta(log n) bits. Arithmetic and boolean
operations on words take constant time, presumably because of
hardware parallelism. Arithmetic on larger integers must be
simulated.
Complexity theorists prefer the bitlevel RAM, but it's rarely used by
anyone else. Almost all integerRAM algorithms are implicitly
developed and analyzed on the wordlevel RAM. Maybe it would be more
accurate to say that most algorithms are analyzed on the unrestricted
unitcost integer RAM, but since the integers they create have only
O(log n) bits, they might as well be using the wordlevel RAM.
Essentially the same ideas as the HartmanisSimonSch\"onhage QBF
algorithm lead to "unreasonably efficient" algorithms and data
structures on the wordlevel RAM, starting with Fredman and Willard's
fusion trees [fwsitbf93].
Anyway... since a unitcost integer RAM can be trivially simulated
using a unitcost real RAM with the floor function, we can solve QBF
or #SAT in polynomial time in that model as well.
The most obvious way to avoid this mess is simply to never combine
exact real arithmetic with the floor function, but it's too late for
that; the bits are already out of the bag. A more reasonable approach
might be to allow the use of the floor function, but only if the
resulting integers have O(log n) bits. Most realRAM+floor algorithms
(that I know about) obey this restriction. Allowing an operation that
computes the O(log n) most significant bits of a real number might
also be reasonable. Alternately, if you prefer the logarithmiccost
model, you could let the cost of the floor operation be the number of
bits of the result.
Even without the floor function, the real RAM is not necessarily a
"reasonable" model of computation. For example, despite the existence
of lineartime real RAM algorithms to compute minimumlink paths,
Kahan and Snoeyink [ksbcmlp96] constructed examples of polygons with
O(log n)bit integer coordinates, such that some minimum link paths
require Omega(n^2 log n) bits of precision. RealRAM algorithms that
compare sums of distances are also questionable, since it is open
whether sums of square roots of integers can be compared in polynomial
time on an integer RAM. (But see [bcsrpt91]!)
 Jeff
(Any references not listed here are in geom.bib.)
@inproceedings{hspmram74
, author = "Juris Hartmanis and Janos Simon"
, title = "On the power of multiplciation in randomaccess machines"
, booktitle = "Proc. 15th Annu. IEEE Sympos. Switching Automata Theory"
, year = 1974
, pages = "1323"
}
@inproceedings{spram79
, author = "Arnold Sch{\"o}nhage"
, title = "On the power of random access machines"
, booktitle = "Proc. 6th Internat. Colloq. Automata Lang. Program."
, series = "Lecture Notes Comput. Sci."
, volume = 71
, publisher = "SpringerVerlag"
, year = 1979
, pages = "520529"
}
@article{bmsscram85
, author = "A. Bertoni and G. Mauri and N. Sabadini"
, title = "Simulations among classes of random access machines and equivalence among numbers succinctly represented"
, journal = "Ann. Discrete Math."
, volume = 25
, year = 1985
, pages = "6590"
}
@incollection{emms90
, author = "Peter van {Emde Boas}"
, title = "Machine models and simulation"
, editor = "Jan van Leeuwen"
, booktitle = "Handbook of Theoretical Computer Science"
, volume = "A"
, publisher = "Elsevier"
, address = "Amsterdam"
, year = 1990
, pages = "166"
}

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From Andreas.Fabri at sophia.inria.fr Mon Dec 8 22:09:23 2003
From: Andreas.Fabri at sophia.inria.fr (andreas)
Date: Mon Jan 9 13:41:13 2006
Subject: Approximation of a function wth piecewise constant functions
MessageID: <3FD4E883.4030505@sophia.inria.fr>
Hello,
I have 100 values between 0 and 255.
I want to approximate this function
with 10 nonoverlapping intervals,
where the degree of freedom is
* the start and endpoint
* the value associated to the interval.
This is probably a classical problem
but I got no reply on sci.opresearch.
Thanks in advance,
andreas
 next part 
An HTML attachment was scrubbed...
URL: http://compgeom.poly.edu/pipermail/compgeomannounce/attachments/20031208/e1b9859f/attachment.htm
From suresh at research.att.com Mon Dec 8 17:23:44 2003
From: suresh at research.att.com (Suresh Venkatasubramanian)
Date: Mon Jan 9 13:41:13 2006
Subject: Approximation of a function wth piecewise constant functions
InReplyTo: <3FD4E883.4030505@sophia.inria.fr>
References: <3FD4E883.4030505@sophia.inria.fr>
MessageID:
On Mon, 8 Dec 2003, andreas wrote:
> Hello,
>
> I have 100 values between 0 and 255.
> I want to approximate this function
> with 10 nonoverlapping intervals,
> where the degree of freedom is
> * the start and endpoint
> * the value associated to the interval.
>
> This is probably a classical problem
> but I got no reply on sci.opresearch.
>
> Thanks in advance,
>
> andreas
>
>
You didn't mention the error function: assuming it is some kind of least
squares, what you want a kmedian solution (k=10 in this case). The
kmedian problem in R^n with metric d is
Find k centers c_1, ... c_k such that the cost
\sum d(p_i, c_n(i)) is minimized, where c_n(i) is the center closest to
p_i.
In your case, you are looking to solve kmedian on the line.
If your error metric is more of an l_infty kind of thing (i.e minimize the
max error), then this boils down to the kcenter problem: same as the
above, except that you want to minimize the max error, rather than the
sum of errors.
There are a variety of algorithms for this problem depending on the
metric, so a search for kmedian/kcenter results should help narrow
things down a bit.
Suresh Venkatasubramanian, Ph: 973 360 8951 (o)
Member, Technical Staff Web: http://www.research.att.com/~suresh/
AT&T Shannon Labs
"The guitar is the ideal instrument for anyone who is able to love
loneliness." Angelo Gilardino

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From rraman at cs.uiowa.edu Tue Dec 9 22:00:38 2003
From: rraman at cs.uiowa.edu (Rajiv Raman)
Date: Mon Jan 9 13:41:13 2006
Subject: Intersecting rectangles.
InReplyTo: <20031207101610.GI9674@granmapa.cs.uiuc.edu>
MessageID:
Hi,
I'm experimenting with some coloring algorithms for intersection graphs of
(isothetic) rectangles (primarily in the plane, but want to experiment
with these techniques for higher dimensional rectangles also).
In this context, rectangles are said to intersect only if the area of
intersection is nonzero.
In order to test these algorithms, I want to write a program that would
take as parameters, (n,k), where n is the number of rectangles, and k is
the maximum number of rectangles that share a point in their interior.
(And hence, the corresponding graph has a clique of size atmost k). The
program would generate n rectangles at random with the desired
intersection property.
I was wondering if there were known efficient algorithms/datastructures
for this problem, or what would be a simple way to implement this.
All techniques I could think of involved generating a random rectangle,
and then testing if it satisifies the intersection constraint. If it
doesn't, then I throw the rectangle away and generate another, and
continue till I have generated n rectangles.
However, this doesn't seem to work well for large values of n and small
values of k.
I would be grateful if anyone could provide pointers in this regard.
Thanks,
Rajiv

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From casa2004 at miralab.unige.ch Wed Dec 10 14:29:29 2003
From: casa2004 at miralab.unige.ch (Casa2004)
Date: Mon Jan 9 13:41:13 2006
Subject: CfP  CASA2004 / Computer Animation and Social Agents 2004
MessageID: <033501c3bf21$a39d8020$a444c281@miralabnt.unige.ch>
[Apologies if you receive this CfP more than once]
 Call for Participation 
 CASA2004 
Computer Animation and Social Agents 2004
http://casa2004.miralab.unige.ch/
 next part 
An HTML attachment was scrubbed...
URL: http://compgeom.poly.edu/pipermail/compgeomannounce/attachments/20031210/c2bd535d/attachment.htm
From bender at cs.sunysb.edu Tue Dec 9 12:36:50 2003
From: bender at cs.sunysb.edu (Michael Bender)
Date: Mon Jan 9 13:41:13 2006
Subject: SPAA '04
MessageID: <200312091736.MAA06319@compserv4.cs.sunysb.edu>
========================================================================
CALL FOR PAPERS
Sixteenth Annual ACM Symposium on
Parallelism in Algorithms and Architectures
SPAA '04
June 2730, 2004
Barcelona, Spain
http://www.spaaconference.org
========================================================================
SCOPE:
Contributed papers are sought in ALL areas of parallel algorithms and
architectures. SPAA defines the term "parallel" broadly, encompassing any
computational system that can perform multiple operations or tasks
simultaneously. Thus, this call for papers covers both traditional
parallel and distributed algorithms and architectures, as well as the
Internet, the web, sensor networks, quantum and DNA computing, etc. Topics
of interest include, but are not limited to:
PARALLEL COMPUTING MASSIVE PARALLELISM
Parallel Algorithms Cluster Computing and Grid Computing
Parallel Complexity Theory Mobile and Wireless Computing
Parallel Computing and Applications The Internet and the World Wide Web
Models for Parallel Sensor Networks
and Distributed Computing
Instruction Level Parallelism and VLSI Satellite and Radio Networks
Routing and Information Dissemination Other Large Networks
Parallel Data Structures Pricing and Equilibria in Networks
Compilers and Tools Supercomputer Architecture
for Parallel Computation and Computing
MemoryAware Algorithms Quantum and DNA Computing
Metacomputing Parallel Data Bases and Data Mining
CONFERENCE PRESENTATION:
Regular presentations will be allotted a 25minute talk and up to 10 pages
in the proceedings. This format is intended for contributions reporting
original research, submitted exclusively to this conference.
Presentation at the SPAA Revue will be allotted a 10minute talk and up to
2 pages in the proceedings. This format is a forum for brief
communications, which may be published later in other conferences.
SUBMISSIONS:
Authors of contributed papers are encouraged to submit their manuscript
electronically. To submit electronically, visit
http://sigact.cs.unlv.edu/~spaa2004/SPAA2004.html for instructions. This
is the preferred method of submission. The deadline for electronic
submissions is February 4, 2004, 5 p.m. EST. The submissions server can be
turned off anytime after this point.
Authors who are unable to submit electronically should contact the program
chair, Micah Adler, at micah@cs.umass.edu to receive instructions. Do not
send electronic submissions to this email address.
The cover page should include (1) title, (2) authors and affiliation,
(3) postal and email address of the contact author, (4) a brief abstract
describing the content of the paper, and (5) an indication if this is a
regular presentation or a SPAA Revue presentation. If requested by the
authors, an extended abstract that is not selected for a regular
presentation will also be considered for the SPAA Revue. Such a request
will not affect the consideration of the paper for a regular presentation.
Submissions for regular presentations should include an introduction
understandable to a nonspecialist including motivation and previous work,
and a technical exposition directed to a specialist. It should not exceed
10 printed pages in 11point type or larger (excluding cover, figures, and
references). More details may be supplied in a clearly marked appendix to
be read at the discretion of the program committee. A cameraready copy
of each accepted paper must be prepared according to ACM guidelines for
inclusion in the conference proceedings.
A submission for the SPAA Revue should consist of a 2page abstract for
each proposed presentation. A cameraready copy of each accepted abstract
will have to be prepared according to ACM guidelines for inclusion in the
proceedings of the conference.
NOTIFICATION
Authors will be sent notification of acceptance or rejection by email or
letter mailed on or before March 15, 2004. A cameraready copy of each
accepted paper, prepared according to ACM guidelines, must be received by
April 6, 2004.
==========================================================================
Program Chair
Micah Adler, U. Massachusetts
Program Committee
Micah Adler, U. Massachusetts
John Byers, Boston U.
Tom Cormen, Dartmouth College
Bruce Hendrickson, Sandia National Laboratories
Maurice Herlihy, Brown U.
Christos Kaklamanis, U. Patras
Christian Lengauer, U. Passau
Geppino Pucci, U. Padova
Satish Rao, U.C. Berkeley
Yves Robert, ENS Lyon
Peter Sanders, MPI Saarbrucken
Daniel Sorin, Duke U.
Aravind Srinivasan, U. Maryland
Berthold Vocking, U. Dortmund
SPAA Local Arrangements Chair
Eulalia Barriere, Technical U. of Catalonia
SPAA General Chair
Phil Gibbons, Intel Research
SPAA Secretary
Cynthia A. Phillips, Sandia National Laboratories
SPAA Treasurer
Rajmohan Rajaraman, Northeastern U.
Publicity Chair
Michael Bender, SUNY Stony Brook

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From shlomo.anglister at intel.com Thu Dec 11 15:26:04 2003
From: shlomo.anglister at intel.com (Anglister, Shlomo)
Date: Mon Jan 9 13:41:13 2006
Subject: Intersecting rectangles.
MessageID: <30B8853201F31F4283E8B901EFE8B5AB029AE711@hasmsx401.iil.intel.com>
Hi,
Consider the following algorithms:
1) Generate a random rectangular cover of the plane.
2) Shift all rectangles at random until you satisfy the clique
condition.
This is a naive approach with bad complexity that will do the job.
1) Generate flowers (cliques) built from rectangles with the desired
clique size.
2) Spread them in the plane.
Hope it helps.
Shlomo
Original Message
From: Rajiv Raman [mailto:rraman@cs.uiowa.edu]
Sent: Wednesday, December 10, 2003 6:01 AM
To: compgeomdiscuss@research.belllabs.com
Subject: Intersecting rectangles.
Hi,
I'm experimenting with some coloring algorithms for intersection graphs
of
(isothetic) rectangles (primarily in the plane, but want to experiment
with these techniques for higher dimensional rectangles also).
In this context, rectangles are said to intersect only if the area of
intersection is nonzero.
In order to test these algorithms, I want to write a program that would
take as parameters, (n,k), where n is the number of rectangles, and k is
the maximum number of rectangles that share a point in their interior.
(And hence, the corresponding graph has a clique of size atmost k). The
program would generate n rectangles at random with the desired
intersection property.
I was wondering if there were known efficient algorithms/datastructures
for this problem, or what would be a simple way to implement this.
All techniques I could think of involved generating a random rectangle,
and then testing if it satisifies the intersection constraint. If it
doesn't, then I throw the rectangle away and generate another, and
continue till I have generated n rectangles.
However, this doesn't seem to work well for large values of n and small
values of k.
I would be grateful if anyone could provide pointers in this regard.
Thanks,
Rajiv

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From tsimos at mail.ariadnet.gr Tue Dec 16 23:10:47 2003
From: tsimos at mail.ariadnet.gr (Theodore Simos)
Date: Mon Jan 9 13:41:13 2006
Subject: ICNAAM 2004
MessageID: <3FDF74D7.FBC1119B@mail.ariadnet.gr>
FIRST ANNOUNCEMENT AND CALL FOR PAPERS
INTERNATIONAL CONFERENCE OF NUMERICAL ANALYSIS AND APPLIED MATHEMATICS
2004
(ICNAAM 2004), CHALKIS , GREECE , 1014 SEPTEMBER 2004.
URL address: http://www.uop.gr/~icnaam/
The aim of ICNAAM 2004 is to bring together leading scientists of the
international Numerical & Applied Mathematics community and to attract
original research papers of very high quality. The topics to be covered
include (but are not limited to): All the research areas of Numerical
Analysis and Computational Mathematics and all the research areas of
Applied Mathematics (see http://www.uop.gr/~icnaam/res8/aimscope.htm).
Chairmen and Organisers
Dr. T.E. Simos, Active Member of the European Academy of Sciences and
Arts and Corresponding Member of the European Academy of Sciences,
Department of Computer Science and Technology, Faculty of Sciences and
Technology, University of Peloponnese, Greece and Dr. Ch. Tsitouras ,
Technological Educational Institute of Chalkis, Greece .
ViceChairman:
Dr. G. Psihoyios, Anglia Polytechnic University , Cambridge , UK .
Scientific Committee
Prof. G. vanden Berghe, Belgium, Prof. P. E. Bjorstad, Norway, Prof. J.
Cash, UK, Prof. R. Cools, Belgium, Prof. A. Cuyt, Belgium, Prof. B.
Fischer, Germany, Prof. R. W. Freund, USA, Prof. I. Gladwell, USA, Prof.
B. Hendrickson, USA, Prof. A. Klar, Germany, Prof. W. F. Mitchell, USA,
Dr. T.E. Simos, Greece, Prof. W.Sproessig, Germany, Dr. Ch. Tsitouras,
Greece, Prof. G. Alistair Watson, UK.
Proceedings: Extended abstracts will be published in a Special Volume
of WileyVCH. The journals in which selected Proceedings of ICNAAM 2004
will be published are: (i) Applied Numerical Analysis and Computational
Mathematics (ANACM) (WileyVCH). This is the official journal of
European Academy of Computational Methods in Sciences and Engineering
and (ii) Mathematical Methods in the Applied Sciences (Wiley & Sons).
Call for Sessions Workshops and Minisymposia: We invite proposals for
Sessions, Workshops or Minisymposia. Each session should have at least 6
paper presentations. For this session the organiser or his team can have
at most 2 papers. Each workshop or minisymposium should have at least 8
paper presentations. For this workshop or minisymposium the organiser or
his team can have at most 2 papers. The Session, Workshop or
Minisymposium organizer will be responsible for advertising the
workshop, reviewing and selecting the papers. The Session organisers
will have free registration in ICNAAM 2004. The Workshop or
Minisymposium organizers will have free registration and a participation
in the Accommodation. Papers accepted for Sessions, Workshops or
Minisymposia will be published in the Proceedings of ICNAAM 2004. After
the Conference the papers presented at the Sessions, Workshops or
Minisymposia will be considered for publication in the appropriate
journals.
Submission of Proposals
Proposals to organize Sessions, Workshops or Minisymposia should include
the following information: Title of the workshop
name, affiliation, mailing address and email address of the proposer(s)
description of the topic of the session (not exceeding 100 words) a
short description on how the session will be advertised. The deadline
for proposal submission is January 15, 2004. Please send your proposal
to icnaam@uop.gr
Contact information: Secretary ICNAAM, Email: icnaam@uop.gr, Postal
Address: 26 Menelaou Street, Amfithea Paleon Faliron, GR175 64, Athens,
Greece, Fax: +30210 94 20 091

Dr. T.E. Simos
Active Member of the European Academy of Sciences and Arts
Corresponding Member of the European Academy of Sciences
EditorinChief and Founder:
Journal of Computational Methods in Sciences and Engineering (JCMSE)
(Cambridge International Sciences Publishing)
Applied Numerical Analysis and Computational Mathematics (ANACM)
(WileyVCH)
Computing Letters (COMPULETT)
(Cambridge International Sciences Publishing)
Editor of the Book Series:
Computational, Numerical and Mathematical Methods in Sciences and
Engineering
(Imperial College Press)

Office Address:
Department of Computer Science and Technology
School of Sciences and Technology
University of Peloponnese
GR22100 Tripolis
Greece

Postal Address
26 Menelaou Street, Amfithea ? Paleon Faliron, GR175 64 Athens, Greece

Conferences:
International Conference of Computational Methods in Sciences and
Engineering 2004 (ICCMSE 2004). We will provide information soon.
For previous International Conferences of Computational Methods in
Sciences and Engineering see at: www.uop.gr/~iccmse/
International Conference of Numerical Analysis and Applied Mathematics
2004 (ICNAAM 2004). URL address: www.uop.gr/~icnaam/

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From vavasis at cs.cornell.edu Wed Dec 17 12:33:36 2003
From: vavasis at cs.cornell.edu (Stephen Vavasis)
Date: Mon Jan 9 13:41:13 2006
Subject: trees inside delaunay triangulation
MessageID: <40E631F174C41E4DBE52727E137AF9277317F2@EXCHVS1.cs.cornell.edu>
Dear colleagues,
I have the following conjecture based on trying some examples. I'm wondering if anyone knows whether it is true.
Conjecture. Let S be a set of n points in general position in the plane. Let T be the Delaunay triangulation of these points. Then there exists a subset T_1,...,T_{n2} of n2 triangles from T such that
(1) These triangles form a tree in the dual graph of the Delaunay triangulation
and
(2) Every one of the n original points is adjacent to at least one of the n2 triangles.
Remark 1. This conjecture is obviously true if the n points are in convex position, since in this case the DT has exact n2 triangles and they form a tree.
Remark 2. The assumption of general position is necessary, since I thought of a counterexample for points not in general position. Specifically, consider a 3by3 evenly spaced rectangular lattice of nine points. There are many DT's for these points, but select the one in which the central point is adjacent to only four edges. This triangulation does not have the conjectured subtree.
Thanks,
Steve Vavasis

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From rseidel at stone.cs.unisb.de Wed Dec 17 21:00:45 2003
From: rseidel at stone.cs.unisb.de (Raimund Seidel)
Date: Mon Jan 9 13:41:13 2006
Subject: trees inside delaunay triangulation
InReplyTo: <40E631F174C41E4DBE52727E137AF9277317F2@EXCHVS1.cs.cornell.edu>
MessageID:
If I am not mistaken, then this is not true in general.
If it were true, the boundary of the tree triangles would
form a hamiltonian circuit.
If I remember correctly Mike Dillencourt showed that there
are Delaunay triangulations that are not Hamiltonian.
RS
On Wed, 17 Dec 2003, Stephen Vavasis wrote:
> Dear colleagues,
>
> I have the following conjecture based on trying some examples. I'm wondering if anyone knows whether it is true.
>
> Conjecture. Let S be a set of n points in general position in the plane. Let T be the Delaunay triangulation of these points. Then there exists a subset T_1,...,T_{n2} of n2 triangles from T such that
> (1) These triangles form a tree in the dual graph of the Delaunay triangulation
> and
> (2) Every one of the n original points is adjacent to at least one of the n2 triangles.
>
> Remark 1. This conjecture is obviously true if the n points are in convex position, since in this case the DT has exact n2 triangles and they form a tree.
>
> Remark 2. The assumption of general position is necessary, since I thought of a counterexample for points not in general position. Specifically, consider a 3by3 evenly spaced rectangular lattice of nine points. There are many DT's for these points, but select the one in which the central point is adjacent to only four edges. This triangulation does not have the conjectured subtree.
>
> Thanks,
> Steve Vavasis
>
> 
> The compgeom mailing lists: see
> http://netlib.belllabs.com/netlib/compgeom/readme.html
> or send mail to compgeomrequest@research.belllabs.com with the line:
> send readme
> Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
>

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From orourke at turing.csc.smith.edu Wed Dec 17 15:26:45 2003
From: orourke at turing.csc.smith.edu (Joseph O'Rourke)
Date: Mon Jan 9 13:41:13 2006
Subject: trees inside delaunay triangulation
InReplyTo:
MessageID:
On Wed, 17 Dec 2003, Raimund Seidel wrote:
> If I am not mistaken, then this is not true in general.
> If it were true, the boundary of the tree triangles would
> form a hamiltonian circuit.
> If I remember correctly Mike Dillencourt showed that there
> are Delaunay triangulations that are not Hamiltonian.
Raimund is correct. I raised this question in an old paper,
and Dillencourt resolved it negatively. References below. :j
@article{obwcdnh87
, author = "J. O'Rourke and H. Booth and R. Washington"
, title = "Connectthedots: {A} new heuristic"
, journal = "Comput. Vision Graph. Image Process."
, volume = 39
, year = 1987
, pages = "258266"
, keywords = "pattern recognition, Delaunay triangulations, Hamiltonian
cycles
"
}
@article{dnhndt87
, author = "M. B. Dillencourt"
, title = "A non{Hamiltonian}, nondegenerate {Delaunay}
triangulation"
, journal = "Inform. Process. Lett."
, volume = 25
, year = 1987
, pages = "149151"
}

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.
From statistics at hicstatistics.org Fri Dec 19 09:20:09 2003
From: statistics at hicstatistics.org (Hawaii International Conference on Statistics)
Date: Mon Jan 9 13:41:13 2006
Subject: Call for Papers
MessageID:
Call for Papers/Abstracts/Submissions
Hawaii International Conference on Statistics, Mathematics and Related
Fields
June 9  12, 2004
Sheraton Waikiki Hotel, Honolulu Hawaii, USA
Submission Deadline: January 21, 2004
Sponsored by:
East West Council for Education
Center of Asian Pacific Studies of Peking University
Web address: http://www.hicstatistics.org
Email address: statistics@hicstatistics.org
The 2004 Hawaii International Conference on Statistics, Mathematics and
Related Fields will be held from June 9 (Wednesday) to June 12 (Saturday),
2004 at the Sheraton Waikiki Hotel in Honolulu, Hawaii. The conference will
provide many opportunities for academicians and professionals from
statistics and/or mathematics related fields to interact with members inside
and outside their own particular disciplines. Crossdisciplinary
submissions with other fields are welcome.
Topic Areas (All Areas of Statistics and/or Mathematics are Invited):
Statistics Topics:
Agricultural Statistics, Applied Statistics, Bayesian Statistics,
Biostatistics, Business Statistics, Computational Statistics, Computer
Simulations, Econometrics, Educational Statistics, Environmental Statistics,
Epidemiology, Industrial Statistics, Management Science, Mathematical
Statistics, Medical Statistics, NonParametric Statistics, Operations
Research, Probability, Psychological Measurement, Quantitative Methods,
Statistical Modeling, Statistics Education, Crossdisciplinary areas of
Statistics, Other Areas of Statistics.
Mathematics Topics:
Algebra, Applied Mathematics, Calculus, Computational Mathematics, Discrete
Mathematics, Foundations of Mathematics, Financial Mathematics, Finite
Mathematics, Fractals, Geometry, History of Mathematics, Logic, Mathematics
Education, Number Analysis, Number Theory, PreAlgebra, PreCalculus,
Probability, Topology, Crossdisciplinary areas of Mathematics, Other Areas
of Mathematics.
The Hawaii International Conference on Statistics, Mathematics and Related
Fields encourages the following types of papers/abstracts/submissions for
any of the listed areas:
Research Papers  Completed papers.
Abstracts  Abstracts of completed or proposed research.
Student Papers  Research by students.
WorkinProgress Reports or Proposals for future projects.
Reports on issues related to teaching.
For more information about submissions see:
http://www.hicstatistics.org/cfp_stats.htm
Format of Presentations:
Paper sessions will have three to four papers presented in each 90 minute
session, giving each presenter 20 ? 30 minutes.
Workshop presentations will be given a full 90 minute session.
Panel sessions will provide an opportunity for three or more presenters to
speak in a more open and conversational setting with conference attendees.
Submissions for these 90 minute sessions should include the name,
department, affiliation, and email address of each panelist in addition to a
description of the presentation and the title page.
Poster sessions will last 90 minutes and consist of a large number of
presenters. Poster sessions allow attendees to speak with the presenters on
a onetoone basis.
Submitting a Proposal:
1. Create a title page for your submission. The title page should include:
a. title of the submission
b. topic area of the submission (chooses from above list)
c. presentation format (choose from above list)
d. name(s) of the author(s)
e. department(s) and affiliation(s)
f. mailing address(es)
g. email address(es)
h. phone number(s)
i. fax number(s)
j. corresponding author if different than lead author
2. Email your abstract and/or paper, along with a title page, to
statistics@hicstatistics.org Receipt of submissions will be acknowledged via
email within 48 hours.
If you do not wish to email your submission, you may send it via regular
mail or fax to:
Hawaii International Conference on Statistics, Mathematics and Related
Fields
P.O. Box 75023
Honolulu, HI, 96836, USA
8089472420 (Fax)
***If submitting via regular mail, please supply two copies of your
submission***
There is a limit of two contributed submissions per lead author.
3. Submissions will only be published in the conference proceedings if at
least one of the authors registers and attends the conference. More
information will be provided upon acceptance.
4. If you wish to be a session chair, please email your request to
statistics@hicstatistics.org and indicate the topic area in which you are
interested. Registration for the conference is required to be a session
chair.
To be removed from this list, please reply to this email with the word
"Remove" in the subject heading.
Hawaii International Conference on Statistics, Mathematics and Related
Fields
P.O. Box 75023
Honolulu, HI 96836 USA
Telephone: (808) 9469927
Fax: (808) 9472420
Email: statistics@hicstatistics.org
Website: www.hicstatistics.org
 next part 
An HTML attachment was scrubbed...
URL: http://compgeom.poly.edu/pipermail/compgeomannounce/attachments/20031219/01acf7c2/attachment.htm
From bender at cs.sunysb.edu Tue Dec 30 12:03:50 2003
From: bender at cs.sunysb.edu (Michael Bender)
Date: Mon Jan 9 13:41:13 2006
Subject: SPAA '04 Extended Deadline
MessageID: <200312301703.MAA19171@compserv4.cs.sunysb.edu>
Please note that the deadline has been extended to February 6, 2004.
It is now after the notification date of STOC.
========================================================================
CALL FOR PAPERS
Sixteenth Annual ACM Symposium on
Parallelism in Algorithms and Architectures
SPAA '04
June 2730, 2004
Barcelona, Spain
http://www.spaaconference.org
========================================================================
SCOPE:
Contributed papers are sought in ALL areas of parallel algorithms and
architectures. SPAA defines the term "parallel" broadly, encompassing any
computational system that can perform multiple operations or tasks
simultaneously. Thus, this call for papers covers both traditional
parallel and distributed algorithms and architectures, as well as the
Internet, the web, sensor networks, quantum and DNA computing, etc. Topics
of interest include, but are not limited to:
PARALLEL COMPUTING MASSIVE PARALLELISM
Parallel Algorithms Cluster Computing and Grid Computing
Parallel Complexity Theory Mobile and Wireless Computing
Parallel Computing and Applications The Internet and the World Wide Web
Models for Parallel Sensor Networks
and Distributed Computing
Instruction Level Parallelism and VLSI Satellite and Radio Networks
Routing and Information Dissemination Other Large Networks
Parallel Data Structures Pricing and Equilibria in Networks
Compilers and Tools Supercomputer Architecture
for Parallel Computation and Computing
MemoryAware Algorithms Quantum and DNA Computing
Metacomputing Parallel Data Bases and Data Mining
CONFERENCE PRESENTATION:
Regular presentations will be allotted a 25minute talk and up to 10 pages
in the proceedings. This format is intended for contributions reporting
original research, submitted exclusively to this conference.
Presentation at the SPAA Revue will be allotted a 10minute talk and up to
2 pages in the proceedings. This format is a forum for brief
communications, which may be published later in other conferences.
SUBMISSIONS:
Authors of contributed papers are encouraged to submit their manuscript
electronically. To submit electronically, visit
http://sigact.cs.unlv.edu/~spaa2004/SPAA2004.html for instructions. This
is the preferred method of submission. The deadline for electronic
submissions is February 6, 2004, 5 p.m. EST. The submissions server can be
turned off anytime after this point.
Authors who are unable to submit electronically should contact the program
chair, Micah Adler, at micah@cs.umass.edu to receive instructions. Do not
send electronic submissions to this email address.
The cover page should include (1) title, (2) authors and affiliation,
(3) postal and email address of the contact author, (4) a brief abstract
describing the content of the paper, and (5) an indication if this is a
regular presentation or a SPAA Revue presentation. If requested by the
authors, an extended abstract that is not selected for a regular
presentation will also be considered for the SPAA Revue. Such a request
will not affect the consideration of the paper for a regular presentation.
Submissions for regular presentations should include an introduction
understandable to a nonspecialist including motivation and previous work,
and a technical exposition directed to a specialist. It should not exceed
10 printed pages in 11point type or larger (excluding cover, figures, and
references). More details may be supplied in a clearly marked appendix to
be read at the discretion of the program committee. A cameraready copy
of each accepted paper must be prepared according to ACM guidelines for
inclusion in the conference proceedings.
A submission for the SPAA Revue should consist of a 2page abstract for
each proposed presentation. A cameraready copy of each accepted abstract
will have to be prepared according to ACM guidelines for inclusion in the
proceedings of the conference.
NOTIFICATION
Authors will be sent notification of acceptance or rejection by email or
letter mailed on or before March 15, 2004. A cameraready copy of each
accepted paper, prepared according to ACM guidelines, must be received by
April 6, 2004.
==========================================================================
Program Chair
Micah Adler, U. Massachusetts
Program Committee
Micah Adler, U. Massachusetts
John Byers, Boston U.
Tom Cormen, Dartmouth College
Bruce Hendrickson, Sandia National Laboratories
Maurice Herlihy, Brown U.
Christos Kaklamanis, U. Patras
Christian Lengauer, U. Passau
Geppino Pucci, U. Padova
Satish Rao, U.C. Berkeley
Yves Robert, ENS Lyon
Peter Sanders, MPI Saarbrucken
Daniel Sorin, Duke U.
Aravind Srinivasan, U. Maryland
Berthold Vocking, U. Dortmund
SPAA Local Arrangements Chair
Eulalia Barriere, Technical U. of Catalonia
SPAA General Chair
Phil Gibbons, Intel Research
SPAA Secretary
Cynthia A. Phillips, Sandia National Laboratories
SPAA Treasurer
Rajmohan Rajaraman, Northeastern U.
Publicity Chair
Michael Bender, SUNY Stony Brook

The compgeom mailing lists: see
http://netlib.belllabs.com/netlib/compgeom/readme.html
or send mail to compgeomrequest@research.belllabs.com with the line:
send readme
Now archived at http://www.uiuc.edu/~sariel/CG/compgeom/maillist.html.