2
$\begingroup$

The question is almost the same as here.

What is the upper bound for a complex Beta function $$\DeclareMathOperator{\Im}{Im}\DeclareMathOperator{\Re}{Re} \displaystyle B(s,z)=\frac{\Gamma(s) \Gamma(z)}{\Gamma(s+z)} $$ where $0<\Re(s)<1$, $0<\Re(z)<1$, $\Im(s)>10$ and $\Im(z)>10$ ?

$\endgroup$
2
  • $\begingroup$ Why the choice of 10? I assumed that you wanted an upper bound or asymptotics as a function of Im(s) and Im(z) $\endgroup$ Commented Oct 18, 2021 at 21:53
  • $\begingroup$ @Yemon Choi values start at ~14. non-trivial zeros of the Riemann zeta function $\endgroup$ Commented Oct 18, 2021 at 22:04

1 Answer 1

0
$\begingroup$

The command of Mathematica

NMaximize[{ComplexExpand[Abs[Gamma[s + I*t]*Gamma[a + I*b]/Gamma[s + a + I*(t + b)]]],s > 0 && a > 0}, {s, t, a, b}] 

outputs "NMaximize::cvdiv: Failed to converge to a solution. The function may be unbounded." and

{4.7857*10^8, {s -> 0.32798, t -> 0.720557, a -> 0., b -> -2.08956*10^-9}}

The command of Mathematica

s = 1/2; t = a; MaxLimit[ ComplexExpand[Abs[Gamma[s + I*t]*Gamma[a + I*b]/Gamma[s + a + I*(t + b)]]], {t,b} -> {0, 0},Direction -> "FromAbove"] 

results in $\infty$, confirming it.

Addition. Here is the answer with Mathematica to the modified question (Thanks to Brendan McKay who paid my attention to the edit.)

NMaximize[{ComplexExpand[ Abs[Gamma[s + I*t]*Gamma[a + I*b]/ Gamma[s + a + I*(t + b)]]], s > 0 && a > 0 && s < 1 && a < 1 && b > 10 && t > 10}, {s, t, a, b}] 

{1.121, {s -> 3.70155*10^-7, t -> 10., a -> 3.57795*10^-7, b -> 10.}}

$\endgroup$
5
  • $\begingroup$ Thanks, but I'm not asking about the limit, but about the absolute value. $\endgroup$ Commented Oct 18, 2021 at 7:13
  • 1
    $\begingroup$ @user363337: The second result means that $$ \left| \frac{\Gamma (a+b i) \Gamma (s+i t)}{\Gamma (a+(b+t) i+s)}\right|$$ is unbounded on the set $a>0,a<1,s>0,s<1$ . Don't hesitate to ask for further explanation in need. $\endgroup$ Commented Oct 18, 2021 at 7:30
  • 2
    $\begingroup$ While I don't know if Mathematica is right or wrong, I don't think the results of executing commands like this form a proof. $\endgroup$ Commented Nov 8, 2023 at 12:59
  • $\begingroup$ @BrendanMcKay: Can you ground your claim? Are you aware of Four colors theorem? $\endgroup$ Commented Nov 8, 2023 at 15:40
  • 2
    $\begingroup$ I use computers to prove things all the time. Symbolic algebra systems make plenty of mistakes, especially when the problem does not have a clean deterministic algorithm such as this 4-dimensional optimisation. Anyway, the OP added lower bounds on the imaginary parts (I think after your answer) that you haven't used. $\endgroup$ Commented Nov 8, 2023 at 23:05

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.