8

In discrete calculus one soon meets the $h$-difference operator $$\Delta_h[f(x)] = f(x+h) - f(x)$$ and we often define $\Delta = \Delta_1.$ We can similarly define the indefinite sums $\Delta_h^{-1}$ and set $\Delta^{-1} = \Delta_1^{-1}.$

Most books on discrete calculus only include results for $h=1.$ For example, the above link gives that $$\Delta^{-1}\sin rx = \frac{-\cos(r[x-\frac{1}{2}])}{2\sin\frac{r}{2}}$$

Through an ugly computation I managed to work out that $$\Delta_h^{-1}[\sin rx] = \frac{-\cos(r[x-\frac{h}{2}])}{2\sin\frac{rh}{2}}$$

What I'd like is a uniform procedure for recovering the "$h$-version" from such results. How could I have derived the second formula from the first? Is there something like the change of variable formula in calculus? Maybe some version of 'dimensional analysis' that tells you where to insert $h$'s? As another example, given that $$\Delta[\log x] = \log(1 + \frac{1}{x})$$ I'd like to be able to see immediately that $$\Delta_h[\log x] = \log(1 + \frac{h}{x})$$

Any help or references would be welcome.

Kariuki
  • 153
  • 2
    I don't know if it's helpful, but if $h\in \Bbb N$ one could write$$f(x+h)-f(x)=[f(x+h)-f(x+h-1)]+[f(x+h-1)-f(x)]=\cdots\=\sum_{n=0}^{h-1}\Delta_1 f(x+h-n)$$ – user170231 Jun 21 '23 at 17:35
  • 2
    have you tried defining $,f_h(x) := f(hx),$ and then define $,g(x) := \Delta_1[f_h(x)],$ which gives $, g(x) = f_h(x+1) - f_h(x) = f(hx+h) - f(hx).,$ Now $, \Delta_h[f(x)] = g(x/h).$ – Somos Jun 23 '23 at 01:24
  • 1
    Could you expand on how this would help me see the correct transformation? – Kariuki Jun 23 '23 at 05:44
  • 2
    crossposted at https://mathoverflow.net/q/449377/11260 – Carlo Beenakker Jun 25 '23 at 12:18

1 Answers1

3

In the context of discrete calculus define the $h$-difference operator

$$\Delta_h[f(x)] := f(x+h) - f(x). $$

Notice that $\,\Delta = \Delta_1\,$ where $\,\Delta\,$ is defined as the difference operator

$$ \Delta[f(x)] := f(x+1) - f(x). $$

You asked the question

What I'd like is a uniform procedure for recovering the "h-version" from such results.

Sadly, the answer, in general, is no. Consider any function that has period $1$. That is, $\,f(x+1) = f(x)\,$ for all $\,x.\,$ This is equivalent to $\,\Delta[f(x)] = 0.\,$ Thus, all information about the function is lost. There is no way to compute $\,\Delta_h[f(x)]\,$ from the $\,\Delta_1\,$ result.

Again, consider any function $\,f(x)\,$ and suppose that the first difference $\,g(x) := \Delta[f(x)]\,$ is known. Then the $2$-difference is $\,\Delta_2[f(x)] = f(x+2)-f(x) = g(x)+g(x+1).\,$ now, in general, these is no easy way to express this in terms of $\,g(x).\,$

This is only a special case of the general problem. Suppose given any function of two variables $\,f(h,x)\,$ where we only know the explicit expression for $\,f(1,x).\,$ It is impossible, in general, to recover the expression for $\,f(h,x).\,$

However, given more information, it may be possible to get results.

For the $h$-difference example, suppose the expression for the function $$ F_r(x) := \Delta[f(rx)] = f(rx+r) - f(rx) $$ is known. Then $$ \Delta_h[f(rx)] = f(r(x+h)) - f(rx) = f(rx+rh) - f(rx) = F_{rh}(x/h). $$

For the indefinite sum example, suppose the expression for the function $$ G_r(x) := \Delta^{-1}[f(rx)] $$ is known where $\,\Delta[G_r(x)] = G_r(x+1) - G_r(x) = f(rx).\,$ Then similarly $$ \Delta_h^{-1}[f(rx)] = G_{rh}(x/h) $$ because $$ \Delta_h[G_{rh}(x/h)] = G_{rh}\left(\frac{x+h}h\right) - G_{rh}\left(\frac{x}h\right) \\ = G_{rh}\left(\frac{x}h+1\right) - G_{rh}\left(\frac{x}h\right) = f(rx). $$

Somos
  • 37,457
  • 3
  • 35
  • 85
  • 1
    I do think this "recovery" becomes possible if you know $\Delta_1[f(rx)]$ for any real $r$ (as in the examples given by the OP); see https://mathoverflow.net/a/449547/11260 – Carlo Beenakker Jun 25 '23 at 12:22
  • 1
    @CarloBeenakker Thanks for that comment! I will use my old comment in my answer. – Somos Jun 25 '23 at 12:41
  • just for the record: my comment was in response to an earlier version of this answer, without the part starting with "However, given more information, it may be possible to get results." – Carlo Beenakker Jun 25 '23 at 13:55