Documente Academic
Documente Profesional
Documente Cultură
2 𝑥 2 2 𝑥3 1 𝑥5 1 𝑥7
In the above, 𝑒𝑟𝑓 = ∫ 𝑒 −𝑡 𝑑𝑡 = √𝜋 (𝑥 − + 2! − 3! + ⋯ ) governs
√𝜋 0 3 5 7
errors’ function widely used in the theory of probability.
This series has been obtained through the decomposition of the integrand
exponential function in accordance with the formula (2.7) and then through
successive integrations. One finally gets
∞
2 (−1)𝑛
𝑓(𝜀) = 1 − ∑ =𝐴
√𝜋 𝑛=0 (2𝑛 + 1)𝑛! 𝜀 2𝑛+1
On the other hand, successive integration of the function (2.47) by parts can
yield an infinite asymptotic series of rhe form
2 ∞
𝜀𝑒 −1⁄𝜋 1 . 3 … . (2𝑛 − 1) 2𝑛
𝑓(𝜀) = [1 + ∑(−1) 𝜀 ] − 𝐵.
√𝜋 2𝑛
𝑛=1
eries B is divergent for arbitrary ɛ. In spite of that, since for small ɛ first terms of
series B decrease quickly, one may obtain more accurate representation of 𝑓(𝜀),
than that achieved through the application of series A. The reason is that the
convergence of series A is very slow. For example, for ɛ = 0.01 one may take
two terms of series B in order to determine 𝑓(𝜀) with a relative error of order
10−8.