|
Computer Sciences Seminar
Thursday, March 7
11:30 AM, NAC 8/206
Evasive Graph Properties and other Hardness Results
Amit Chakrabarti
Princeton University
Abstract
Computational complexity theory seeks to understand the capabilities
as well as the limitations of algorithms. Hardness results, also
called lower bound results, are aimed at the latter goal. Given our
very limited success at proving hardness results in general
computational models (such as the Turing machine), research in this
area has tended to focus on more specialized models, tailored to the
problem at hand.
A long-standing family of open problems in this area concerns algorithms
that decide whether or not an input graph satisfies a particular
property. The most celebrated of these open problems is Karp's
"evasiveness conjecture", which states that for a monotone graph
property any deterministic algorithm must inspect every single edge of
the graph, in the worst case. Examples of monotone properties are
connectedness, k-colorability, Hamiltonicity, and just about any other
"natural" graph property.
We prove that Karp's conjecture is true for the important class of
minor-closed graph properties. We also significantly improve earlier
results for subgraph containment properties. I shall present these
results, along with a brief survey of my other work on computational
hardness.
Bio
Amit Chakrabarti received a B.Tech. in Computer Science from the Indian
Institute of Technology, Bombay, along with the President of India Gold
Medal. He subsequently received an M.A. in Computer Science and expects
to receive a Ph.D. in Computer Science from Princeton University in
summer 2002. His research interests are in Complexity Theory, where he
investigates the limitations of computers at solving fundamental
algorithmic problems, and in Approximation Algorithms, where he is
interested in finding close-to-optimal solutions for NP-hard
optimization problems.
|