1

Say I want to define a function fwith input x but also involving parameter p not yet specified:

f[x_] := p^x

Now I want to verify some inequalities with additional assumptions. A (hopefully) minimal example, showing my problem:

If, for some for input y in the function and for some paremeter value q>0, it holds that f[y] >= 2 q, then I would expect f[y] >= q to yield True.

However,

FullSimplify[f[y] >= q, {f[y] >= 2 q, q > 0}]

yields p^y >= q instead. Even if I add q \[Element] Reals as an assumption.

Do I overlook an obvious mathematical (non-Mathematica) issue, or how I can get Mathematica to "fully" simplify the expression? I have also looked at this issue with FullSimplify, but the solution provided did not help.

Bernd
  • 965
  • 9
  • 19
  • Have you tried the functions ForAll and Resolve? – Mr Puh May 13 '20 at 10:56
  • I rather seek to verify whether for some specific (not all/arbitrary) values of the parameters certain statements hold true. So as in the example, I have a general function f, "local" assumption f[y]>=2q with q>0 and the statement I want to verify, f[y] >= q, which I expected to turn into True. ForAll and Resolve don't seem to fare any bettr. I hoped my example was minimalistic enough, I can add more if required. – Bernd May 13 '20 at 11:21
  • I am not sure how capable Mathematica is in solving these kind of analytical questions. It might also be due to the complexity of your function f. Try using f[x_] := Hold[E^(Log[p] x)] as your equation, this seems to give the desired result. – Mr Puh May 13 '20 at 12:07

1 Answers1

1

I am not certain that this is what you are looking for?

f[x_] := p^x
Reduce[f[x] > q && q > 0, p, Reals]

Visual:

With[
     {f = p^x},
     Manipulate[
      Plot[{f, q}, {x, 0, 2}, Filling -> {1 -> {{2}, {Orange, Blue}}}],
      {{p, 1/2}, -1, 1},
      {{q, 1/2}, -4, 4}
      ]
     ]
Craig Carter
  • 4,416
  • 16
  • 28
  • I am not so much interested in the requirements/ranges for the parameters for certain (in)equalities to be true (which, I presume, Reduce would give me to some extend), but rather in having local restrictions/assumptions on certain values, then trying to check how these simplify my computations. I.e. "if some statement x holds, can I then infer y"? So I would want to "reduce" w.r.t. all parameters, i.e. some sort of "simplifying". – Bernd May 13 '20 at 10:28