Jump to content

Talk:Time complexity

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Mistake

[edit]

There is a mistake in description of Quasi-polynomial time: 3sat it is not NP hard problem but NP complete, we have a polynomial verification for the solution of the problem.

NPC is the intersection of NP and NPH, so all NPC problems are NPH. See NP-hardness to learn more. (I did fix an unrelated error, though.) 2620:0:1A10:7816:4CE9:BD3B:BE35:4850 (talk) 22:30, 15 December 2021 (UTC)[reply]

It wasn't an error, though. Polynomial and sublinear time algorithms are still also quasipolynomial time algorithms. Your edit introduced an error by saying that they are not. —David Eppstein (talk) 01:07, 16 December 2021 (UTC)[reply]

APR-CL time complexity

[edit]

In the "Superpolynomial time" section the APR algorithm is given as an example for algorithm with time complexity of , but in the article about the algorithm the complexity mentioned is . This might be caused if the complexity in the other article uses to represent the unary length of the number instead of binary but I don't think that is the case. —YotamEdit (talk) 18:51, 7 July 2021 (UTC)[reply]

P and NP Problems.

[edit]

My question is: If N^3 is p type problem, and to solve this it's take almost 3 hours, Then now I want to ask what is value of N^3. 2402:8100:2652:B42D:61B8:A047:8958:1CFB (talk) 07:51, 29 March 2022 (UTC)[reply]

What is T(n)?

[edit]

In the section on 'Constant Time', the function T(n) first shows up without any previous explanation. There should be a section differentiating b/w Time Complexity and 'Order/degree of Time Complexity'. — Preceding unsigned comment added by Granzer92 (talkcontribs) 07:33, 26 May 2022 (UTC)[reply]

Article too long

[edit]

This article is too long. I suggest splitting it into separate articles. At the very least, polynomial time should have its own article. Other important complexity classes should have their own articles, too. 2601:547:B05:ECD:ED93:4837:4D47:671 (talk) 06:29, 31 January 2023 (UTC)[reply]

At 33kB "readable prose size" we are at the point where WP:TOOBIG says "Length alone does not justify division". But some of these time classes may well work better as standalone topics. The bigger problem is that many sections are completely unsourced. With sources, they could stand up a lot better to expansion and splitting off into separate articles. Without sources, we cannot make articles out of them. —David Eppstein (talk) 07:58, 31 January 2023 (UTC)[reply]

Is factorial too narrow?

[edit]

Is "factorial" a too narrow class? The other at-least-subexponential classes mentioned here are closed under multiplication of the exponent by a constant, but this definition of factorial isn't. Factorial is between 2^((n/2)log(n/2)) and 2^(n log n), so maybe 2^O(n log n) is a more useful definition, more in the spirit of other at-least-subexponential classes, and captures similar functions like n^n. The examples mention Bogosort, which is O(n*n!), which is actually not O(n!), but is 2^O(n log n).

SvenSandberg (talk) 07:08, 14 October 2023 (UTC)[reply]