19.4: Stirling's Approximation
- Page ID
- 151784
The polynomial coefficient, \(C\), is a function of the factorials of large numbers. Since \(N!\) quickly becomes very large as \(N\) increases, it is often impractical to evaluate \(N!\) from the definition,
\[N!=\left(N\right)\left(N-1\right)\left(N-2\right)\dots \left(3\right)\left(2\right)\left(1\right) \nonumber \]
Fortunately, an approximation, known as Stirling’s formula or Stirling’s approximation is available. Stirling’s approximation is a product of factors. Depending on the application and the required accuracy, one or two of these factors can often be taken as unity. Stirling’s approximation is
\[N!\approx N^N \left(2\pi N\right)^{1/2}\mathrm{exp}\left(-N\right)\mathrm{exp}\left(\frac{1}{12N}\right)\approx N^N\left(2\pi N\right)^{1/2}\mathrm{exp}\left(-N\right)\approx N^N\mathrm{exp}\left(-N\right) \nonumber \]
In many statistical thermodynamic arguments, the important quantity is the natural logarithm of \(N!\) or its derivative, \({d ~ { \ln N!\ }}/{dN}\). In such cases, the last version of Stirling’s approximation is usually adequate, even though it affords a rather poor approximation for \(N!\) itself.