Hans-Joachim Böckenhauer: Catalogue data in Spring Semester 2024 |
| Name | Dr. Hans-Joachim Böckenhauer |
| Consultation hours | By appointment |
| Address | Professur Algorithmen und Didaktik ETH Zürich, CAB F 11 Universitätstrasse 6 8092 Zürich SWITZERLAND |
| hjb@inf.ethz.ch | |
| URL | http://www.ite.ethz.ch/people/hjb/ |
| Department | Computer Science |
| Relationship | Lecturer |
| Number | Title | ECTS | Hours | Lecturers | ||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 252-4910-00L | Algorithms with Predictions   The deadline for deregistering expires at the end of the second week of the semester. Students who are still registered after that date, but do not attend the seminar, will officially fail the seminar. | 2 credits | 2S | H.‑J. Böckenhauer, F. Frei, D. Komm | ||||||||||||||||||||||||||||||||
| Abstract | Online algorithms have to answer a sequence of requests without knowing the whole sequence beforehand. We want to survey recent results about how predictions about future requests, given, e.g., by some machine-learning approach, can influence the performance of online algorithms. | |||||||||||||||||||||||||||||||||||
| Learning objective | To systematically acquire an overview of the impact of side information on the performance of online algorithms, especially in the context of , e.g., machine-learned, predictions about future requests. | |||||||||||||||||||||||||||||||||||
| Content | In the classical model of online algorithms, one assumes that the input is revealed piecewise in the form of requests over time and an algorithm has to respond with a part of the output to each request. While there are many situations in which this model is more realistic than the classical model of computation where the whole input is known in advance, not knowing anything about future requests is a quite pessimistic assumption. Recently, several approaches have been introduced to incorporate some kind of predictions about future requests into the model. These predictions can, e,g,, be based on some statistical knowledge about typical instances or can be generated by some machine-learning approaches. In this seminar, after some brief introduction to online algorithms, we want to explore the impact of different kinds of predictions on the solution quality of online algorithms. Each participant will study one aspect of this topic, following a specific scientific publication, and will give a presentation about this topic. | |||||||||||||||||||||||||||||||||||
| Literature | The literature will consist of textbook chapters and original research papers and will be provided during the kick-off meeting. | |||||||||||||||||||||||||||||||||||
| Prerequisites / Notice | The participants should be familiar with the content of the lectures "Algorithmen und Datenstrukturen" (252-0026-00) and "Theoretische Informatik" (252-0057-00). The presentations will be given in the form of a block course in June 2024, preferably shortly after the end of the normal lectures. The language can be mixed in German and English in the following sense: The teaching material will be in English, but it will be possible for the participants to give their presentations and hand in their written summaries in German. | |||||||||||||||||||||||||||||||||||
| Competencies |
| |||||||||||||||||||||||||||||||||||
| 272-0302-00L | Approximation and Online Algorithms | 5 credits | 2V + 1U + 1A | H.‑J. Böckenhauer, D. Komm, M. Wettstein | ||||||||||||||||||||||||||||||||
| Abstract | This lecture deals with approximative algorithms for hard optimization problems and algorithmic approaches for solving online problems as well as the limits of these approaches. | |||||||||||||||||||||||||||||||||||
| Learning objective | Get a systematic overview of different methods for designing approximative algorithms for hard optimization problems and online problems. Get to know methods for showing the limitations of these approaches. | |||||||||||||||||||||||||||||||||||
| Content | Approximation algorithms are one of the most succesful techniques to attack hard optimization problems. Here, we study the so-called approximation ratio, i.e., the ratio of the cost of the computed approximating solution and an optimal one (which is not computable efficiently). For an online problem, the whole instance is not known in advance, but it arrives pieceweise and for every such piece a corresponding part of the definite output must be given. The quality of an algorithm for such an online problem is measured by the competitive ratio, i.e., the ratio of the cost of the computed solution and the cost of an optimal solution that could be given if the whole input was known in advance. The contents of this lecture are - the classification of optimization problems by the reachable approximation ratio, - systematic methods to design approximation algorithms (e.g., greedy strategies, dynamic programming, linear programming relaxation), - methods to show non-approximability, - classic online problem like paging or scheduling problems and corresponding algorithms, - randomized online algorithms, - the design and analysis principles for online algorithms, and - limits of the competitive ratio and the advice complexity as a way to do a deeper analysis of the complexity of online problems. | |||||||||||||||||||||||||||||||||||
| Literature | The lecture is based on the following books: J. Hromkovic: Algorithmics for Hard Problems, Springer, 2004 D. Komm: An Introduction to Online Computation: Determinism, Randomization, Advice, Springer, 2016 Additional literature: A. Borodin, R. El-Yaniv: Online Computation and Competitive Analysis, Cambridge University Press, 1998 | |||||||||||||||||||||||||||||||||||
| Competencies |
| |||||||||||||||||||||||||||||||||||