The design of approximation algorithms is an important
theme in the study of optimization problems.
It takes on an additional significance when applied
to NP-hard problems.
We describe a highly efficient scheme
for converting a pseudo epsilon-approximation
algorithm into a true epsilon-approximation algorithm.
The advantage for this approach derives from
(i) the relative ease of constructing pseudo approximation
algorithms as compared to approximation algorithms,
(ii) the use of the above generic conversion scheme,
and (iii) a clearer understanding of those aspects of the approximation
process that are sensitive to the precision of the input.
We illustrate our technique on two problems:
(A) Euclidean Shortest Path (3ESP), namely
the shortest path for a point robot amidst polyhedral obstacles in 3-D, and
(B) d1-optimal motion for a rod moving amidst
polygonal obstacles in 2-D.
These problems are the simplest NP-hard motion
planning problems in 3-D and 2-D, respectively.
Previously, no epsilon-approximation algorithm
for (B) was known. For (A), our new solution
is simpler than two previous solutions and also has
an exponentially lower complexity
in terms of the input precision parameter.
Joint work with Tetsuo Asano and David Kirkpatrick.
Thursday, March 28 at 4pm in DH 3076
A 3:30 reception in DH 3076 will precede the talk.
About Chee Yap
Chee Yap is Professor of Computer Science at the
Courant Institute of Mathematical Sciences, New York University.
He has published over 120 papers in major conferences and journals,
in the area of algorithms and complexity.
His research interests include computational geometry, visualization problems,
computer algebra and algorithmic robotics.
Since 1993, he has been interested in numerical non-robustness,
especially issues at the interface between numerical, algebraic and
topological computation. A current project called the Core Library
(http://cs.nyu.edu/exact/core/) aims to bring robust computation out of
research realm and make it widely accessible to all programmers.
Yap was born in Singapore and educated at MIT (S.B. 1975) and Yale (PhD 1980).
He served on the editorial boards of SIAM J. of Computing (1985-2002),
J. of Computer and System Sciences (1986--),
J. of Symbolic Computation (1988--),
Computational Geometry: Theory and Applications (1990--),
Int'l J. of Comp. Geometry and Applications (1990--),
and Algorithmica (1992--). His books include the edited volume
"Advances in Robotics: Algorithmic and Geometric Issues",
(co-edited with Jack Schwartz), Lawrence Erlbaum Associates, Inc, 1987,
and "Fundamental Problems in Algorithmic Algebra",
Oxford University Press, 2000.