Conference item icon

Conference item

Shortest paths without a map, but with an entropic regularizer

Abstract:
In a 1989 paper titled 'shortest paths without a map', Papadimitriou and Yannakakis introduced an online model of searching in a weighted layered graph for a target node, while attempting to minimize the total length of the path traversed by the searcher. This problem, later called layered graph traversal, is parametrized by the maximum cardinality k of a layer of the input graph. It is an online setting for dynamic programming, and it is known to be a rather general and fundamental model of online computing, which includes as special cases other acclaimed models. The deterministic competitive ratio for this problem was soon discovered to be exponential in k, and it is now nearly resolved: it lies between O(2k) and O(k2k). Regarding the randomized competitive ratio, in 1993 Ramesh proved, surprisingly, that this ratio has to be at least O(k2/log1+?k) (for any constant ? > 0). In the same paper, Ramesh also gave an O(k13)-competitive randomized online algorithm. Since 1993, no progress has been reported on the randomized competitive ratio of layered graph traversal. In this work we show how to apply the mirror descent framework on a carefully selected evolving metric space, and obtain an O(k2) competitive randomized online algorithm, nearly matching the known lower bound on the randomized competitive ratio.
Publication status:
Published
Peer review status:
Peer reviewed

Actions


Access Document


Files:
Publisher copy:
10.1109/FOCS54457.2022.00036

Authors


More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Computer Science
Oxford college:
St Anne's College
Role:
Author
ORCID:
0000-0003-3744-0977


Publisher:
IEEE
Host title:
Proceedings of the IEEE 63RD Annual Symposium on Foundations of Computer Science (FOCS 2022)
Pages:
1102-1113
Publication date:
2022-12-28
Event title:
IEEE 63RD Annual Symposium on Foundations of Computer Science (FOCS 2022)
Event location:
Denver, CO, USA
Event website:
https://focs2022.eecs.berkeley.edu/
Event start date:
2022-10-31
Event end date:
2022-11-03
DOI:
EISSN:
2575-8454
ISSN:
1523-8288
ISBN:
9781665455190


Language:
English
Keywords:
Pubs id:
1326742
Local pid:
pubs:1326742
Deposit date:
2023-03-28

Terms of use



Views and Downloads






If you are the owner of this record, you can report an update to it here: Report update to this record

TO TOP