October
31- November 1, 2002

Princeton University

Princeton, New Jersey

Schedule and abstracts

Facility location problems were first considered by the Operations
Research community in the context of locating warehouses to minimize
the cost of product distribution. More recent work has demonstrated
a wide variety of applications for these problems. For example,
facility location can be used to model data clustering, an essential
operation in data mining, by replacing the distance metric with
a measurement of object similarity. As another example, problems
in design of network topologies can be viewed as hierarchical
applications of a facility location paradigm. Each of these applications
has given rise to similar but slightly different versions of the
problem. Furthermore, many different approaches have been taken
to producing solutions to these problems. To some degree these
differences reflect distinct problem objectives, but in many cases
they are also a reflection of a lack of communication between
researchers working on similar problems arising from different
fields of study.

In recent years, there has been a surge of interest in facility
location problems in the computer science community, perhaps in
part due to the above-mentioned applications to the design of
computer networks and in clustering of large databases. Computer
science theorists have generally viewed these problems from the
perspective of worst-case performance-guarantee approximation
algorithms, and have produced a number

of such combinatorial solutions for the problem.

This workshop will bring together researchers from Operations
Research, Data Mining, Computer Networking, and Computer Science
in an effort to discuss and better understand the various forms
of facility location problems, and to share useful techniques
in approaching them. We hope to place particular emphasis in discussing
the integration of facility location models and transportation
network design models that have traditionally decoupled in the
process of logistic design. We aim to formulate new meaningful
models at this intersection and stimulate more work in this area.

The workshop will be a combination of survey talks, new results,
and informal discussion. There will be no published proceedings,
but we plan to have a web page for the workshop with slides, pointers
to relevant papers, and so forth.

Organizing committee: Moses
Charikar, Ted Giford, Adam Meyerson,
R. Ravi

See
the PROBE

Schedule
and abstracts

Adam
Meyerson's description of all the talks and results/open problems (ppt presentation)