1 #LyX 2.4 created this file. For more info see https://www.lyx.org/
5 \save_transient_properties true
8 \use_default_options true
9 \maintain_unincluded_children false
11 \language_package default
14 \font_roman "default" "default"
15 \font_sans "default" "default"
16 \font_typewriter "default" "default"
17 \font_math "auto" "auto"
18 \font_default_family default
19 \use_non_tex_fonts false
23 \font_typewriter_osf false
24 \font_sf_scale 100 100
25 \font_tt_scale 100 100
27 \use_dash_ligatures true
29 \default_output_format default
31 \bibtex_command default
32 \index_command default
33 \paperfontsize default
37 \use_package amsmath 1
38 \use_package amssymb 1
41 \use_package mathdots 1
42 \use_package mathtools 1
44 \use_package stackrel 1
45 \use_package stmaryrd 1
46 \use_package undertilde 1
48 \cite_engine_type default
51 \paperorientation portrait
63 \paragraph_separation indent
64 \paragraph_indentation default
66 \math_numbering_side default
71 \paperpagestyle default
73 \tracking_changes false
82 \begin_layout Standard
83 \begin_inset FormulaMacro
84 \newcommand{\ud}{\mathrm{d}}
90 \Delta_{n}(x,z)\equiv\int_{x}^{\infty}t^{-\frac{1}{2}-n}e^{-t+\frac{z^{2}}{4t}}\ud t\label{eq:Delta definition}
98 \begin_layout Standard
99 Integration per partes:
102 \begin_layout Standard
105 \int t^{-\frac{1}{2}-n}\ud t=\frac{t^{\frac{1}{2}-n}}{\frac{1}{2}-n};
113 \frac{\ud}{\ud t}e^{-t+\frac{z^{2}}{4t}}=\left(-1-\frac{z^{2}}{4t^{2}}\right)e^{-t+\frac{z^{2}}{4t}}
121 \begin_layout Standard
124 \left(\frac{1}{2}-n\right)\Delta_{n} & =-x^{\frac{1}{2}-n}e^{-x+\frac{z^{2}}{4x}}+\int_{x}^{\infty}t^{\frac{1}{2}-n}e^{-t+\frac{z^{2}}{4t}}\ud t+\frac{z^{2}}{4}\int_{x}^{\infty}t^{\frac{-3}{2}-n}e^{-t+\frac{z^{2}}{4t}}\ud t\\
125 & =-x^{\frac{1}{2}-n}e^{-x+\frac{z^{2}}{4x}}+\Delta_{n-1}+\frac{z^{2}}{4}\Delta_{n+1},
133 \Delta_{n+1}=\frac{4}{z^{2}}\left(\left(\frac{1}{2}-n\right)\Delta_{n}-\Delta_{n-1}+x^{\frac{1}{2}-n}e^{-x+\frac{z^{2}}{4x}}\right).\label{eq:Delta recurrence}
138 There are obviously wrong signs in Kambe II, (A 3.3).
141 \begin_layout Standard
144 \begin_inset CommandInset ref
146 reference "eq:Delta recurrence"
153 is obviously unsuitable for numerical computation when
154 \begin_inset Formula $z$
158 However, the definition
159 \begin_inset CommandInset ref
161 reference "eq:Delta definition"
168 suggests that the function should be analytical around
169 \begin_inset Formula $z=0$
174 \begin_inset Formula $z=0$
177 , one has (by definition of incomplete Г function)
180 \Delta_{n}(x,0)=\Gamma\left(\frac{1}{2}-n,x\right).\label{eq:Delta:z = 0}
185 For convenience, label
186 \begin_inset Formula $w=z^{2}/4$
192 \Delta'_{n}\left(x,w\right)\equiv\int_{x}^{\infty}t^{-\frac{1}{2}-n}e^{-t+\frac{w}{t}}\ud t.
197 Differentiating by parameter
198 \begin_inset Formula $w$
201 (which should be fine as long as the integration contour does not go through
205 \frac{\partial\Delta'_{n}\left(x,w\right)}{\partial w}=\Delta'_{n+1}\left(x,w\right),
213 \frac{\partial^{k}}{\partial w^{k}}\Delta'_{n}\left(x,w\right)=\Delta'_{n+k}\left(x,w\right).
219 \begin_inset CommandInset ref
221 reference "eq:Delta:z = 0"
228 , this gives an expansion around
229 \begin_inset Formula $w=0$
235 \Delta_{n}'\left(x,w\right)=\sum_{k=0}^{\infty}\Gamma\left(\frac{1}{2}-n-k,x\right)\frac{w^{k}}{k!},
243 \Delta_{n}\left(x,z\right)=\sum_{k=0}^{\infty}\Gamma\left(\frac{1}{2}-n-k,x\right)\frac{\left(z/2\right)^{2k}}{k!}.
248 The big negative first arguments in incomplete
249 \begin_inset Formula $\Gamma$
252 functions should be good (at least I think so, CHECKME), as well as the
254 \begin_inset Formula $1/k!$
258 I am not sure what the convergence radius is, but for
259 \begin_inset Formula $\left|z\right|<2$
262 there seems to be absolutely no problem in using this formula.