Undergrad Smoothness of Discrete Functions

  • Thread starter Thread starter The Number Juggler
  • Start date Start date
  • Tags Tags
    Discrete Functions
Click For Summary
Techniques for measuring the smoothness of discrete functions involve concepts similar to those used for differentiable functions, such as bounded first and partial derivatives. The discussion highlights the importance of understanding the function's domain, whether it involves a single variable or multiple variables. The "calculus of finite differences" is suggested as a relevant method for analyzing smoothness. Specific criteria, such as those from actuarial science, define smoothness in terms of differences and growth rates. Overall, the conversation emphasizes the need for established terminology and research on this topic.
The Number Juggler
Messages
1
Reaction score
0
Hi Physics Forums

Is there a specific technique to measure how smooth a discrete function is?

By smooth I mean that if you change the input by a minimum amount then you know that the objective function result will not have a big jump.

For example The Closest String Problem is completely smooth function, since if you change the input string by only one letter, then the distance to the furthest string will change by at most one.

The Traveling Salesman Problem is also smooth, but not completely so. If you change only the order that you visit two cities then the total distance traveled could not change as much as if you altered more. However there could be larger and smaller jumps.

Moving around a grid of random numbers would be completely unsmooth.

I'm sure there must be another word for this phenomena and some research done already. But google is providing anything except "smoothing discrete functions" which is a completely different thing altogether!
 
Physics news on Phys.org
The Number Juggler said:
Hi Physics Forums

Is there a specific technique to measure how smooth a discrete function is?

By smooth I mean that if you change the input by a minimum amount then you know that the objective function result will not have a big jump.

It isn't clear what you have in mind for the domain of the function - one real variable ?, a vector of variables ?

If you want to look at methods that are analogous to those used for differentiable functions (e.g. first derivative being bounded or partial derivatives being bounded) investigate "the calculus of finite differences".
 
The only thing I can recall is a function is from actuarial science: Barnett n-th order smooth iff ##|{\Delta^n a_x}| \cdot 7^n < a_x##. But this is really for life tables.
 
A discrete function is smooth if f(2n) = Θ(n) (theta of n). If f(2n) = o(n), it is not smooth. Source: Algorithms by Neapolitan, Appendix B.2.
 
The standard _A " operator" maps a Null Hypothesis Ho into a decision set { Do not reject:=1 and reject :=0}. In this sense ( HA)_A , makes no sense. Since H0, HA aren't exhaustive, can we find an alternative operator, _A' , so that ( H_A)_A' makes sense? Isn't Pearson Neyman related to this? Hope I'm making sense. Edit: I was motivated by a superficial similarity of the idea with double transposition of matrices M, with ## (M^{T})^{T}=M##, and just wanted to see if it made sense to talk...

Similar threads

  • · Replies 7 ·
Replies
7
Views
4K
  • · Replies 6 ·
Replies
6
Views
3K
  • · Replies 0 ·
Replies
0
Views
5K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
1
Views
3K
  • · Replies 125 ·
5
Replies
125
Views
20K
  • · Replies 8 ·
Replies
8
Views
4K
Replies
2
Views
4K
Replies
7
Views
3K