Can I publish an algorithm without time complexity?

Question

There's a problem 'A' and there are already algorithms to solve it but I came up with a totally new algorithm. My doubts are:

Should I find time complexity?
Where can I publish my algorithm without time complexity?
Should I publish an algorithm in a math or computer science engineering journal?

Strongly related question: Is improving a simple algorithm without beating the state of the art still publishable? — Nobody, Aug 17 '18 at 11:27
Without knowing the time complexity of an algorithm, it's hard to say anything about it. Analyze the complexity of your algorithm before trying to publish. — Bob Brown, Aug 17 '18 at 12:39
It is not even sure that time complexity is still relevant. It is mostly for sequential algorithms, but most of gains in hardware power these days come from adding more cores and relates to parallellizeable algorithms. — mathreadler, Aug 19 '18 at 05:34
@mathreadler Time complexity is still a thing for parallel algorithms, usually stated as a function of the input size and the number of processors. — JeffE, Aug 20 '18 at 14:25
@JeffE yees yees of course it is a thing, but how it scales with nr of processors becomes more and more and more important. — mathreadler, Aug 20 '18 at 14:46
@mathreadler That scaling is precisely what "time complexity" means for parallel algorithms. — JeffE, Aug 20 '18 at 16:46

score 4 · Answer 1 · answered Aug 17 '18 at 11:38

An algorithm that solves a problem in a new way is interesting, regardless of whether you beat the state-of-the-art, because that new way might start a new line of research that beats the state-of-the-art. But, as it stands, you seem unaware of where your algorithm ranks and establishing that is highly desirable (perhaps even necessary) prior to publication. Thus, although it might be possible to publish now, it seems better to delay publication to establish your algorithm's ranking.

score 2 · Answer 2 · answered Aug 17 '18 at 11:49

If you submit your paper to a journal it will face review. If it passes review, the editors will consider it for publication. However, I strongly suspect that the reviewers will question the lack of the analysis in their report and ask for it to be included. That isn't necessarily the case, but the paper/algorithm otherwise needs to be interesting enough for them to want to see it in print even without the analysis. I doubt that will happen, however.

You are free to submit, but there are a couple of gates you must pass through on the way to publication. Better to make the passage as easy as possible.

Can I publish an algorithm without time complexity?

2 Answers2