### DIMACS Workshop on Probabilistic Analysis of Algorithms

#### May 11-14, 1997

Princeton University

**Organizers:**
**Alan Frieze**, Carnegie Mellon, af1p+@andrew.cmu.edu
**Michael Molloy**, University of Toronto, molloy@cs.toronto.edu

Presented under the auspices of the DIMACS Special Focus on Discrete Probability.

### Theme:

Problems arising in the mathematical
analysis of algorithms have provided an important stimulus
for the growth in interest in Discrete Mathematics.
Probabilistic techniques have an important role in this endeavour. The
theory of algorithms has
undergone an extraordinarily vigorous development over the last
20 years or so, and probability theory has emerged as one of its most
vital partners.
By its nature probabilistic analysis cuts across the boundaries of
Computer Science, Discrete Mathematics,
Operations Research and Probability Theory.
This workshop focusses on the analysis of algorithms in the {\em average
case}, assuming
a probabilistic distribution on input instances. The goal is to obtain good
estimates of various performance measures of algorithms such as solution
quality and running time.
The main objective is
to explain why many simple algorithms, often used in practice,
perform much better than as predicted by worst-case analysis.
Pathological cases
are generally rare and simple algorithms can often be shown to perform
well in the average case. This is especially true in the case of
NP-complete problems
where it seems that the
probabilistic approach may be the best way
and indeed may be the only way to understand the success of
heuristic combinatorial algorithms.

We identify five major sub-areas as the focus of the workshop.

- Algorithmic Theory of Random Graphs and Related Combinatorial Problems.
- Analysis of Euclidean Functionals.
- Linear Programming and Integer Programming Problems.
- Analysis of Fundamental Algorithms.
- Average Case Completeness

The workshop will bring together leading researchers in the Analysis of
Algorithms along
with younger researchers in the area. There should be about 15-20
lectures presenting
progress, trends and open problems together with ample time for
discussion. We believe that
bringing together researchers from these different but closely related
areas will result in
an interesting cross-fertilisation of ideas.

Next: Call for Participation

Index

DIMACS Homepage

Contacting the Center

Document last modified on November 2, 1998.