10.11.09
|
Isolde Adler (Goethe-Universität Frankfurt am Main)
Graphzerlegungen und Model Checking
Abstract:
Ist derzeit nicht verfügbar.
|
24.11.09
|
Stefan C. Ionescu (Goethe-Universität Frankfurt am Main)
Lokale Suche in variabler Tiefe
Abstract:
Die lokale Suche in variabler Tiefe, von Kernighan und Lin vorgestellt,
ist eines der erfolgreichsten Paradigmen für die schnelle Bestimmung guter
Lösungen für das Graph Partitionierung und das Traveling Salesman Problem.
Der für das Problem der Graph Partitionierung 1970
vorgestellte Algorithmus und seine 1973 vorgestellte Adaption für das
Traveling Salesman Problem sind die Basis diverser Varianten und Verbesserungen.
Obwohl die Laufzeiten der verschiedenen, auf dem Kernighan-Lin Paradigma
basierenden Algorithmen im Detail untersucht wurden, so blieb die
Approximationsleistung der beiden Verfahren bisher ungeklärt.
In diesem Vortrag stellen wir das Paradigma der lokalen Suche in
variable Tiefe vor und besprechen abgeleitete Heuristiken und Erweiterungen.
Desweiteren konstruieren wir planare Graphen mit N Knoten und
Approximationsfaktor \Omega(N) für das Problem der Graphpartitionierung.
Die Analyse belegt. dass der Kernighan-Lin Algorithmus für eine große
Klasse von Graphen, beschrieben durch einige wenige Eigenschaften
ebenfalls nur eine sehr schwache Approximationsleistung aufweist.
Unsere Ergebnisse gelten bei zufälliger Wahl der Startzerlegung mit
extrem hoher Wahrscheinlichkeit obwohl sogar intelligentes Tie-Breaking
erlaubt ist.
Abschließend werfen wir noch einen kurzen Blick auf empirische Ergebnisse,
um die Schärfe der analytischen Schranke zu betonen.
|
08.12.09
|
Andrei Negoescu (Goethe-Universität Frankfurt am Main)
Online Paging for Flash Memory Devices
Abstract:
We propose a variation of online paging in two-level memory systems where
pages in the fast cache get modified and therefore have to be
explicitly written back to the slow memory upon evictions. For
increased performance, up to $\alpha$ arbitrary pages can be moved
from the cache
to the slow memory within a single joint eviction, whereas fetching
pages from the slow
memory is still performed on a one-by-one basis. The main objective in
this
new $\alpha$-paging scenario is to bound the number of evictions.
After providing experimental evidence that $\alpha$-paging can
adequately model
flash-memory devices in the context of translation layers
we turn to the theoretical connections between $\alpha$-paging and
standard paging.
We give lower bounds for deterministic and randomized
$\alpha$-paging algorithms. For deterministic algorithms, we show that
an adaptation of LRU is strongly competitive, while for the
randomized case we show that by adapting the classical Mark
algorithm we get an algorithm with a competitive ratio larger
than the lower bound by a multiplicative factor of approximately 1.7.
|
26.01.10
|
Nicole Schweikardt (Goethe-Universität Frankfurt am Main)
Additions-invariantes FO, reguläre Sprachen und semi-lineare Mengen
Abstract:
We consider first-order formulas which, in addition to the symbols
in the vocabulary, may
use two designated symbols < and + that must be interpreted as a linear
order and its associated addition. Such a formula is called
addition-invariant if, for each fixed interpretation of the initial vocabulary,
its result is independent of
the particular interpretation of < and +.
We study the expressive power of addition-invariant first-order
logic, +-inv-FO, on the class of finite strings.
Our first main result gives a characterization of the regular languages
definable in +-inv-FO: we show that these are exactly the languages
definable in FO with extra predicates, denoted by " lm" for short,
for testing the length of the string modulo some fixed number.
Our second main result shows that every language definable in +-inv-FO, that
is bounded or commutative or deterministic context-free, is regular. As an
immediate consequence of these two main results, we obtain that +-inv-FO is
equivalent to FO( lm) on the class of finite colored sets.
Our proof methods involve Ehrenfeucht-Fraissé games,
tools from algebraic automata theory, and reasoning about semi-linear sets.
[This is joint work with Luc Segoufin.]
|
02.02.10
|
Annamaria Kovacs (Goethe-Universität Frankfurt am Main)
Deterministic truthful PTAS for scheduling related machines
Abstract:
Scheduling on related machines (Q||C_max) is one of the most important problems
in the field of Algorithmic Mechanism Design. Each machine is
controlled by a selfish agent and her valuation can be expressed via
a single parameter, her speed.
Archer and Tardos (FOCS01) showed that, in contrast to other similar problems,
a (non-polynomial) allocation that minimizes the makespan can be truthfully implemented.
On the other hand, if we leave out the game-theoretic issues, the complexity of the problem
has been completely settled --- the problem is strongly NP-hard, while there exists a PTAS.
This problem is the most well-studied in single-parameter Algorithmic
Mechanism Design. It gives an excellent ground to explore the boundary between truthfulness
and efficient computation. Since the work of Archer and Tardos, quite a lot of
deterministic and randomized mechanisms have been suggested.
Recently, a breakthrough result of Dhangwatnotai et al. (FOCS08) showed that
a randomized, truthful-in-expectation PTAS exists. On the other hand, for the
deterministic case, the best known approximation factor is 2.8.
It has been a major open question whether
there exists a deterministic truthful PTAS, or whether truthfulness
has an essential, negative impact on the computational complexity of the problem. We
give a definitive answer to this important question by providing a truthful deterministic PTAS.
|
17.02.10
|
Morteza Monemizadeh (TU Dortmund)
Der Vortrag findet am Mittwoch, 17.02.10, 14-16 Uhr statt.
1-Pass Relative-Error L_p-Sampling with Applications
Abstract:
For any p in [0,2], we give a 1-pass poly(\eps^{-1}\log n)-space algorithm which, given a
data stream of length m with insertions and deletions of an n-dimensional vector a, with
updates in the range { -M, -M+1, ... , M-1, M }, outputs a sample of {1,2,... ,n}
for which for all i the probability that i is returned is (1 \pm \epsilon)\frac{|a_i|^p}{F_p(a)} \pm n^{-C},
where a_i denotes the (possibly negative) value of coordinate i, F_p(a) = \sum_{i=1}^n |a_i|^p = ||a||_p^p
denotes the p-th frequency moment (i.e., the p-th power of the L_p norm), and C > 0 is
an arbitrarily large constant. Ignoring an initial preprocessing stage, the update time of the
algorithm is \poly(\eps^{-1}\log n). Here we assume that n,m, and M are polynomially
related.
Our generic sampling framework improves and unifies algorithms for several communication and streaming
problems,
including cascaded norms, heavy hitters, and moment estimation. It also gives the first relative-error
forward
sampling algorithm in a data stream with deletions, answering an open question of
Cormode et al.
At the end of the talk I will also give an overview of the results for various problems
that we obtained
in data streaming model in the last 4 years. This includes new streaming algorithms for
k-means,
subspace approximation, facility location,and regression problems.
[This is joint work with David P. Woodruff, IBM Almaden Research Center, San Jose, CA]
|