Digital Library of Mathematical Functions
About the Project
NIST
12 Parabolic Cylinder FunctionsProperties

§12.2 Differential Equations

Contents

§12.2(i) Introduction

PCFs are solutions of the differential equation

12.2.1\frac{{d}^{2}w}{{dz}^{2}}+\left(az^{2}+bz+c\right)w=0,

with three distinct standard forms

Each of these equations is transformable into the others. Standard solutions are \mathop{U\/}\nolimits\!\left(a,\pm z\right), \mathop{V\/}\nolimits\!\left(a,\pm z\right), \mathop{\overline{U}\/}\nolimits\!\left(a,\pm x\right) (not complex conjugate), \mathop{U\/}\nolimits\!\left(-a,\pm iz\right) for (12.2.2); \mathop{W\/}\nolimits\!\left(a,\pm x\right) for (12.2.3); \mathop{D_{{\nu}}\/}\nolimits\!\left(\pm z\right) for (12.2.4), where

All solutions are entire functions of z and entire functions of a or \nu.

For real values of z (=x), numerically satisfactory pairs of solutions (§2.7(iv)) of (12.2.2) are \mathop{U\/}\nolimits\!\left(a,x\right) and \mathop{V\/}\nolimits\!\left(a,x\right) when x is positive, or \mathop{U\/}\nolimits\!\left(a,-x\right) and \mathop{V\/}\nolimits\!\left(a,-x\right) when x is negative. For (12.2.3) \mathop{W\/}\nolimits\!\left(a,x\right) and \mathop{W\/}\nolimits\!\left(a,-x\right) comprise a numerically satisfactory pair, for all x\in\Real. The solutions \mathop{W\/}\nolimits\!\left(a,\pm x\right) are treated in §12.14.

In \Complex, for j=0,1,2,3, \mathop{U\/}\nolimits\!\left((-1)^{{j-1}}a,(-i)^{{j-1}}z\right) and \mathop{U\/}\nolimits\!\left((-1)^{j}a,(-i)^{j}z\right) comprise a numerically satisfactory pair of solutions in the half-plane \tfrac{1}{4}(2j-3)\pi\leq\mathop{\mathrm{ph}\/}\nolimits z\leq\tfrac{1}{4}(2j+%
1)\pi.

§12.2(ii) Values at z=0

12.2.6\mathop{U\/}\nolimits\!\left(a,0\right)=\frac{\sqrt{\pi}}{2^{{\frac{1}{2}a+%
\frac{1}{4}}}\mathop{\Gamma\/}\nolimits\!\left(\frac{3}{4}+\frac{1}{2}a\right)},
12.2.9{\mathop{V\/}\nolimits^{{\prime}}}\!\left(a,0\right)=\frac{\pi 2^{{\frac{1}{2}%
a+\frac{3}{4}}}}{\left(\mathop{\Gamma\/}\nolimits\!\left(\frac{1}{4}-\frac{1}{%
2}a\right)\right)^{2}\mathop{\Gamma\/}\nolimits\!\left(\frac{3}{4}+\frac{1}{2}%
a\right)}.

§12.2(iv) Reflection Formulas

§12.2(v) Connection Formulas

§12.2(vi) Solution \mathop{\overline{U}\/}\nolimits\!\left(a,x\right); Modulus and Phase Functions

When z (=x) is real the solution \mathop{\overline{U}\/}\nolimits\!\left(a,x\right) is defined by

unless a=\tfrac{1}{2},\tfrac{3}{2},\dots, in which case \mathop{\overline{U}\/}\nolimits\!\left(a,x\right) is undefined. Its importance is that when a is negative and |a| is large, \mathop{U\/}\nolimits\!\left(a,x\right) and \mathop{\overline{U}\/}\nolimits\!\left(a,x\right) asymptotically have the same envelope (modulus) and are \tfrac{1}{2}\pi out of phase in the oscillatory interval -2\sqrt{-a}<x<2\sqrt{-a}. Properties of \mathop{\overline{U}\/}\nolimits\!\left(a,x\right) follow immediately from those of \mathop{V\/}\nolimits\!\left(a,x\right) via (12.2.21).

where F(a,x) (>0), \theta(a,x), G(a,x) (>0), and \psi(a,x) are real. F or G is the modulus and \theta or \psi is the corresponding phase.

For properties of the modulus and phase functions, including differential equations, see Miller (1955, pp. 72–73). For graphs of the modulus functions see §12.3(i).