TY - GEN
T1 - The Wrong Direction of Jensen’s Inequality Is Algorithmically Right
AU - Zamir, Or
N1 - Publisher Copyright:
© Or Zamir.
PY - 2023/7
Y1 - 2023/7
N2 - Let A be an algorithm with expected running time eX, conditioned on the value of some random variable X. We construct an algorithm A′ with expected running time O(eE[X]), that fully executes A. In particular, an algorithm whose running time is a random variable T can be converted to one with expected running time O(eE[ln T]), which is never worse than O(E[T]).
AB - Let A be an algorithm with expected running time eX, conditioned on the value of some random variable X. We construct an algorithm A′ with expected running time O(eE[X]), that fully executes A. In particular, an algorithm whose running time is a random variable T can be converted to one with expected running time O(eE[ln T]), which is never worse than O(E[T]).
KW - Jensen’s inequality
KW - algorithms
KW - complexity
UR - http://www.scopus.com/inward/record.url?scp=85167334555&partnerID=8YFLogxK
U2 - 10.4230/LIPIcs.ICALP.2023.107
DO - 10.4230/LIPIcs.ICALP.2023.107
M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???
AN - SCOPUS:85167334555
T3 - Leibniz International Proceedings in Informatics, LIPIcs
BT - 50th International Colloquium on Automata, Languages, and Programming, ICALP 2023
A2 - Etessami, Kousha
A2 - Feige, Uriel
A2 - Puppis, Gabriele
PB - Schloss Dagstuhl- Leibniz-Zentrum fur Informatik GmbH, Dagstuhl Publishing
T2 - 50th International Colloquium on Automata, Languages, and Programming, ICALP 2023
Y2 - 10 July 2023 through 14 July 2023
ER -