|
|
|
"Mathematical
Challenges in Telecommunication"
|
|
|
|
Martin
Gr\"otschel
|
|
Konrad-Zuse-Zentrum
f\"ur Informationstechnik
|
|
and Technische
Universit\"at Berlin, Germany
|
|
|
|
Abstract:
|
|
|
|
This talk will
begin with a survey of mathematical
|
|
challenges that
arise in telecommunication. Mathematics
|
|
is involved, e.
g., in the design and manufacturing of
|
|
chips, devices
and network components, the choice
|
|
of locations,
the planning of the network topology,
|
|
and the
dimensioning of the equipment involved.
|
|
Adequate
cryptography, the need of fast data processing,
|
|
demand routing
and failure handling require efficient
|
|
and reliable
mathematical algorithms on the operational
|
|
side.
|
|
|
|
The presentation
will focus on the problem of
|
|
designing
low-cost telecommunication networks that
|
|
provide
sufficient capacity to serve a given demand,
|
|
are based on a
chosen technology mix, satisfy various
|
|
technical side
constraints, and survive certain
|
|
failure
situations. This problem is difficult
|
|
in theory and
practice. It will be indicated how
|
|
algorithms
integrating polyhedral combinatorics, linear
|
|
and integer
programming, and various heuristic ideas can
|
|
help solve
real-world instances within reasonable quality
|
|
guarantees in
acceptable running times.
|
|
|
|
The lecture will
also address issues such as: balancing
|
|
the load of
signaling transfer points, issues arising
|
|
in packet
switching, modeling optical switches and
|
|
all optical
networks.networks
|
|
|
|
This talk is
based on work of the telecommunications
|
|
research group
at ZIB, the examples discussed and
|
|
the
computational results reported are from joint
|
|
projects with
several telecommunication companies.
|