Kvantu algoritmi datu plūsmu apstrādei
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
lav
Abstract
Datu plūsma ir informācijas kopums, kas tiek pārraidīts no datu avota noteiktā secībā. Tiek pētīta kvantu algoritmu sarežģītība datu plūsmu modelī ar vairākām plūsmas lasīšanām. Darbā piedāvāti četri kvantu algoritmi datu plūsmu apstrādei, kas pārspēj labākos klasiskos risinājumus. Pirmais algoritms rēķina multigrafa virsotnes pakāpi, izmanto O(log n log(1/ϵ)) atmiņas un lasa plūsmu O(1/ϵ) reizes. Otrais algoritms rēķina multigrafa frekvenču momentu M2, izmanto O((log n)^2+(log(1/ϵ))^2) atmiņas un lasa plūsmu Õ(√n/ϵ^2) reizes. Divi pārējie algoritmi nosaka unikālo elementu skaitu: viens ar atmiņu O(1/ϵ^2 log n) un divām plūsmas lasīšanām, otrs - izmanto mazāk atmiņas O(log n + log(1/ϵ)) un lasa plūsmu Õ(1/ϵ) reizes.
A data stream is a sequence of data transmitted from a source in a specific order. This work studies the complexity of quantum algorithms in the data stream model with multiple passes. Four quantum algorithms for data stream processing are proposed, all outperforming the best-known classical solutions. The first algorithm computes the degree of a vertex in a multigraph, using O(log n log(1/ϵ)) memory and O(1/ϵ) passes. The second algorithm calculates the second frequency moment M2 of the multigraph, using O((log n)^2+(log(1/ϵ))^2) memory and Õ(√n/ϵ^2) passes. The other two algorithms estimate the number of unique elements: one uses O(1/ϵ^2 log n) memory and two passes over the stream, while the other uses less memory, O(log n + log(1/ϵ)), and makes Õ(1/ϵ) passes over the stream.
A data stream is a sequence of data transmitted from a source in a specific order. This work studies the complexity of quantum algorithms in the data stream model with multiple passes. Four quantum algorithms for data stream processing are proposed, all outperforming the best-known classical solutions. The first algorithm computes the degree of a vertex in a multigraph, using O(log n log(1/ϵ)) memory and O(1/ϵ) passes. The second algorithm calculates the second frequency moment M2 of the multigraph, using O((log n)^2+(log(1/ϵ))^2) memory and Õ(√n/ϵ^2) passes. The other two algorithms estimate the number of unique elements: one uses O(1/ϵ^2 log n) memory and two passes over the stream, while the other uses less memory, O(log n + log(1/ϵ)), and makes Õ(1/ϵ) passes over the stream.