 Participants:
Maia Averett, University of California,
Santa Barbara.
Brendan Creutz, Cal Poly, San Luis Obispo.
Patricia Romano Cirilo, Universidade Federal de Minas Gerais.
John Adrian DeIonno, University of California, Los Angeles.
Lisa Helene Feigenbaum, Harvard University.
Juliana Abrantes Freire, PUC  Rio de Janeiro.
Jean Carlo P. Garcia, Universidade Federal do Rio Grande do Sul.
William Jeck, Pomona College.
Mikhail Lev, University of California, Los Angeles.
Ives Jose A. Macedo Jr., Universidade Federal de Pernambuco.
Rafael Kaufmann Nedal, PUC  Rio de Janeiro.
Anna Shustrova, University of California, Berkeley.
Renato R. V. Zanforlin, Universidade Federal de Minas Gerais.
 Brazilian participants were supported by the Brazilian Federal Agency,
CNPq.
 Organizers and Faculty Advisors: M. Helena Noronha,
California State University Northridge, and Carlos Tomei, PUC
 Rio de Janeiro.
 Other Faculty Advisors: Helio Lopes, Marco Grivet,
and Carlos Frederico Borges Palmeira, PUC  Rio de Janeiro.

The focus of the 2003 program was Applied Mathematics.
In the first week participants were introduced to possible research
projects and were given the option to select the problem to work on.
Towards the end of the program, students gave presentations describing
their progress. They also wrote a preliminary paper describing their
results. As the program was only one month long, participants didn't
have much time to revise and polish their papers. These papers were
seen as work in progress and after the summer, students continued to
work with their advisors to prepare papers for publication.
Below are the participants and faculty advisors. Click on the title
of the papers for a pdf version.



Tetrahedral meshes used in applications such as volume visualization
consume very large amounts of memory. Thus compression is essential
for storage and transmission purposes. A lot of different schemes have
been developed to address this problem. Our algorithm handles tetrahedral
meshes that are connected oriented combinatorial 3manifolds without
boundaries. It extends the Edgebreaker algorithm on a Corner Table for
triangular meshes to work with tetrahedral meshes.


The purpose is to minimize the cost of operating
a hydrothermal power system throughout a certain amount of time, while
meeting market demand at every instant. We have considered three models
and applied some mathematical programming techniques.


The goal of this study is to develop a global sense
of how functions from the plane to the plane behave. In the case of
functions from the real line to the real line, the standard practice
is to compute critical points (i.e. maxima and minima) then use these
points to infer geometric information about the function as a whole.
We extend this approach to functions from the plane to the plane. Using
topological tools we combine local theory of the function at singularities
and regular points to obtain a global picture of how the function behaves.
The image of the critical set and the locations of its
preimages are highly structured. This gives us detailed information
concerning the number and location of all preimages of any given point.
In particular, we provide a description of a method for finding the
roots of such functions, which may be applied to solving solutions of
two equations in two unknowns. We then compare the method described
herein for obtaining roots to methods employed by such standard mathematical
packages as Maple and Mathematica.

MultiChannel Wireless Telecommunication
Systems: An Algorithm for Optimal Channel and Power Allocation
Maia Averett, Lisa Feigenbaum, Juliana Freire, and Mikhail Lev.

Optimization of multichannel wireless communication
networks is a field of research, which strives to improve system capacity
by appropriately allocating network resources. This overarching goal
entails a balance between a number of individual goals, including: minimization
of total power consumption, minimization of any individual's power consumption,
and maintenance of a sufficient number of freely available channels.
One current problem in the wireless telecom industry is the lack of
an efficient algorithm for partitioning a set of links into subsets
which can be grouped into various, shared communication resources. We
have chosen to focus on two of the many aspects of this problem. The
first is to find a quick way of determining whether or not a set of
links can share a communication resource feasibility. The second is
to create an efficient algorithm for exploring the possible link combinations
and recording those of which are feasible. We use undergraduate linear
algebra techniques to present and modify parts of an existing algorithm
by Professor Marco Grivet, PUCRio.

