Chebyshev theta function: Difference between revisions
mNo edit summary |
m →Estimates of the function: I didn't know if the slight refinement was due to Chebyshev |
||
| Line 16: | Line 16: | ||
'''Theorem (Chebyshev).''' If <math>x \ge 0</math>, then <math>\vartheta(x) \le | '''Theorem (Chebyshev).''' If <math>x \ge 0</math>, then <math>\vartheta(x) \le | ||
2x</math>. | 2x \log 2</math>. | ||
''Proof.'' We induct on <math>\lfloor x \rfloor</math>. For our base | ''Proof.'' We induct on <math>\lfloor x \rfloor</math>. For our base | ||
cases, we note that for <math>0 \le x < 2</math>, we have <math>\vartheta(x) = | cases, we note that for <math>0 \le x < 2</math>, we have <math>\vartheta(x) = | ||
0 \le | 0 \le 2x \log 2</math>. | ||
Now suppose that <math>x \ge 2</math>. Let <math>n = \lfloor x \rfloor</math>. Then | Now suppose that <math>x \ge 2</math>. Let <math>n = \lfloor x \rfloor</math>. Then | ||
| Line 26: | Line 26: | ||
\prod_{\lfloor n/2 \rfloor < p \le n} p , </cmath> | \prod_{\lfloor n/2 \rfloor < p \le n} p , </cmath> | ||
so | so | ||
<cmath> | <cmath> x \log 2 \ge \sum_{\lfloor n/2 \rfloor < p \le n} \log p | ||
= \vartheta{x} - \vartheta{\lfloor n/2 \rfloor} | = \vartheta{x} - \vartheta{\lfloor n/2 \rfloor} | ||
\ge \vartheta{x} - 2\lfloor n/2 \rfloor \ge \vartheta{x} - x , </cmath> | \ge \vartheta{x} - 2\lfloor n/2 \rfloor \log 2 \ge \vartheta{x} - x \log 2x , </cmath> | ||
by inductive hypothesis. Therefore | by inductive hypothesis. Therefore | ||
<cmath> 2x \ge \vartheta(x), </cmath> | <cmath> 2x \log 2 \ge \vartheta(x), </cmath> | ||
as desired. <math>\blacksquare</math> | as desired. <math>\blacksquare</math> | ||
Revision as of 22:43, 29 March 2009
Chebyshev's theta function, denoted
or sometimes
, is a function of use in analytic number theory.
It is defined thus, for real
:
where the sum ranges over all primes less than
.
Estimates of the function
The function
is asymptotically equivalent to
(the prime counting function) and
. This result
is the Prime Number Theorem, and all known proofs are rather
involved.
However, we can obtain a simpler bound on
.
Theorem (Chebyshev). If
, then
.
Proof. We induct on
. For our base
cases, we note that for
, we have
.
Now suppose that
. Let
. Then
so
by inductive hypothesis. Therefore
as desired.