University of Leicester
Browse

Query Minimization under Stochastic Uncertainty

conference contribution
posted on 2020-05-21, 14:43 authored by Steven Chaplick, Magnus M Halldorsson, Murilo de Lima, Tigran Tonoyan

We study problems with stochastic uncertainty data on intervals for which the precise value can be queried by paying a cost. The goal is to devise an adaptive decision tree to find a correct solution to the problem in consideration while minimizing the expected total query cost. We show that sorting in this scenario can be performed in polynomial time, while finding the data item with minimum value seems to be hard. This contradicts intuition, since the minimum problem is easier both in the online setting with adversarial inputs and in the offline verification setting. However, the stochastic assumption can be leveraged to beat both deterministic and randomized approximation lower bounds for the online setting. Although some literature has been devoted to minimizing query/probing costs when solving uncertainty problems with stochastic input, none of them have considered the setting we describe. Our approach is closer to the study of query-competitive algorithms, and it gives a better perspective on the impact of the stochastic assumption

Funding

Partially supported by Icelandic Research Fund grant 174484-051 and by EPSRC grant EP/S033483/1. This work started while M.S.L. and T.T. were at Reykjavik University, during a research visit by S.C.

History

Citation

Lecture Notes in Computer Science (2020) In Press

Source

14th Latin American Theoretical Informatics Symposium

Version

  • AM (Accepted Manuscript)

Published in

Lecture Notes in Computer Science

Publisher

Springer Verlag

issn

0302-9743

Acceptance date

2020-02-11

Copyright date

2020

Spatial coverage

São Paulo, Brazil

Temporal coverage: start date

2020-05-25

Temporal coverage: end date

2020-05-29

Language

en

Usage metrics

    University of Leicester Publications

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC