Difference between revisions of "Integral logarithm"
Ulf Rehmann (talk | contribs) m (tex encoded by computer) |
m (fix tex) |
||
Line 57: | Line 57: | ||
\mathop{\rm li} ( x) = c | \mathop{\rm li} ( x) = c | ||
+ \mathop{\rm ln} | \mathop{\rm ln} x | + | + \mathop{\rm ln} | \mathop{\rm ln} x | + | ||
− | \sum _ { k= } | + | \sum _ { k= 1} ^ \infty |
\frac{( \mathop{\rm ln} x ) ^ {k} }{k ! k } | \frac{( \mathop{\rm ln} x ) ^ {k} }{k ! k } | ||
Line 71: | Line 71: | ||
\mathop{\rm li} ( z) = c + | \mathop{\rm li} ( z) = c + | ||
\mathop{\rm ln} ( - \mathop{\rm ln} z ) + | \mathop{\rm ln} ( - \mathop{\rm ln} z ) + | ||
− | \sum _ { k= } | + | \sum _ { k=1 } ^ \infty |
\frac{( \mathop{\rm ln} z ) ^ {k} }{k ! k } | \frac{( \mathop{\rm ln} z ) ^ {k} }{k ! k } | ||
Line 88: | Line 88: | ||
$$ | $$ | ||
\lim\limits _ {\eta \downarrow 0 } \mathop{\rm li} ( x \pm i \eta ) | \lim\limits _ {\eta \downarrow 0 } \mathop{\rm li} ( x \pm i \eta ) | ||
− | = \mathop{\rm li} x \ | + | = \mathop{\rm li} x \mp \pi i ,\ \ |
x > 1 . | x > 1 . | ||
$$ | $$ | ||
Line 131: | Line 131: | ||
the value $ \mathop{\rm li} ( x) $ | the value $ \mathop{\rm li} ( x) $ | ||
is a good approximation of $ \pi ( x) $, | is a good approximation of $ \pi ( x) $, | ||
− | the number of primes smaller than $ x $( | + | the number of primes smaller than $ x $ |
− | see [[De la Vallée-Poussin theorem|de la Vallée-Poussin theorem]]; [[Distribution of prime numbers|Distribution of prime numbers]]; [[Prime number|Prime number]]). | + | (see [[De la Vallée-Poussin theorem|de la Vallée-Poussin theorem]]; [[Distribution of prime numbers|Distribution of prime numbers]]; [[Prime number|Prime number]]). |
Latest revision as of 18:28, 25 February 2021
The special function defined, for positive real $ x $,
$ x \neq 1 $,
by
$$ \mathop{\rm li} ( x) = \ \int\limits _ { 0 } ^ { x } \frac{dt}{ \mathop{\rm ln} t } ; $$
for $ x > 1 $ the integrand has at $ t = 1 $ an infinite discontinuity and the integral logarithm is taken to be the principal value
$$ \mathop{\rm li} ( x) = \ \lim\limits _ {\epsilon \downarrow 0 } \ \left \{ \int\limits _ { 0 } ^ { {1 } - \epsilon } \frac{dt}{ \mathop{\rm ln} t } + \int\limits _ {1 + \epsilon } ^ { x } \frac{dt}{ \mathop{\rm ln} t } \right \} . $$
The graph of the integral logarithm is given in the article Integral exponential function. For $ x $ small:
$$ \mathop{\rm li} ( x) \approx \frac{x}{ \mathop{\rm ln} ( 1 / x ) } . $$
The integral logarithm has for positive real $ x $ the series representation
$$ \mathop{\rm li} ( x) = c + \mathop{\rm ln} | \mathop{\rm ln} x | + \sum _ { k= 1} ^ \infty \frac{( \mathop{\rm ln} x ) ^ {k} }{k ! k } ,\ \ k > 0 ,\ \ x \neq 1 , $$
where $ c = 0.5772 \dots $ is the Euler constant. As a function of the complex variable $ z $,
$$ \mathop{\rm li} ( z) = c + \mathop{\rm ln} ( - \mathop{\rm ln} z ) + \sum _ { k=1 } ^ \infty \frac{( \mathop{\rm ln} z ) ^ {k} }{k ! k } $$
is a single-valued analytic function in the complex $ z $- plane with slits along the real axis from $ - \infty $ to 0 and from 1 to $ + \infty $( the imaginary part of the logarithms is taken within the limits $ - \pi $ and $ \pi $). The behaviour of $ \mathop{\rm li} x $ along $ ( 1 , + \infty ) $ is described by
$$ \lim\limits _ {\eta \downarrow 0 } \mathop{\rm li} ( x \pm i \eta ) = \mathop{\rm li} x \mp \pi i ,\ \ x > 1 . $$
The integral logarithm is related to the integral exponential function $ \mathop{\rm Ei} ( x) $ by
$$ \mathop{\rm li} ( x) = \ \mathop{\rm Ei} ( \mathop{\rm ln} x ) ,\ \ x < 1 ; \ \ \mathop{\rm Ei} ( x) = \ \mathop{\rm li} ( e ^ {x} ) ,\ \ x < 0 . $$
For real $ x > 0 $ one sometimes uses the notation
$$ \mathop{\rm Li} ( x) = \ \left \{ \begin{array}{ll} \mathop{\rm li} ( x) = \mathop{\rm Ei} ( \mathop{\rm ln} x ) &\textrm{ for } 0 < x < 1 , \\ \mathop{\rm li} ( x) + \pi i = \mathop{\rm Ei} ^ {*} ( \mathop{\rm ln} x ) &\textrm{ for } x > 1 . \\ \end{array} \right .$$
For references, see Integral cosine.
Comments
The function $ \mathop{\rm li} $ is better known as the logarithmic integral. It can, of course, be defined by the integral (as above) for $ z \in \mathbf C \setminus \{ {x \in \mathbf R } : {x \leq 0 \textrm{ or } x \geq 1 } \} $.
The series representation for positive $ x $, $ x \neq 1 $, is then also said to define the modified logarithmic integral, and is the boundary value of $ \mathop{\rm li} ( x + i \eta ) \pm \pi i $, $ x > 1 $, $ \eta \rightarrow 0 $. For real $ x > 1 $ the value $ \mathop{\rm li} ( x) $ is a good approximation of $ \pi ( x) $, the number of primes smaller than $ x $ (see de la Vallée-Poussin theorem; Distribution of prime numbers; Prime number).
Integral logarithm. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Integral_logarithm&oldid=47376