11 Struve and Related FunctionsRelated Functions

§11.11 Asymptotic Expansions of Anger–Weber Functions

Contents
  1. §11.11(i) Large |z|, Fixed ν
  2. §11.11(ii) Large |ν|, Fixed z
  3. §11.11(iii) Large ν, Fixed z/ν

§11.11(i) Large |z|, Fixed ν

Let F0⁢(ν)=G0⁢(ν)=1, and for k=1,2,3,…,

11.11.1 Fk⁢(ν) =(ν2−12)⁢(ν2−32)⁢⋯⁢(ν2−(2⁢k−1)2)=(−4)k⁢(12−ν2)k⁢(12+ν2)k,
Gk⁢(ν) =(ν2−22)⁢(ν2−42)⁢⋯⁢(ν2−(2⁢k)2)=(−4)k⁢(1−ν2)k⁢(1+ν2)k.

Then as z→∞ in |ph⁡z|≤π−δ

11.11.2 𝐉ν⁡(z)∼Jν⁡(z)+sin⁡(π⁢ν)π⁢z⁢(∑k=0∞Fk⁢(ν)z2⁢k−νz⁢∑k=0∞Gk⁢(ν)z2⁢k),
11.11.3 𝐄ν⁡(z)∼−Yν⁡(z)−1+cos⁡(π⁢ν)π⁢z⁢∑k=0∞Fk⁢(ν)z2⁢k−ν⁢(1−cos⁡(π⁢ν))π⁢z2⁢∑k=0∞Gk⁢(ν)z2⁢k,

For sharp error bounds and exponentially-improved extensions, see Nemes (2018).

§11.11(ii) Large |ν|, Fixed z

If z is fixed, and ν→∞ in |ph⁡ν|≤π in such a way that ν is bounded away from the set of all integers, then

11.11.5 𝐉ν⁡(z)=sin⁡(π⁢ν)π⁢ν⁢(1−ν⁢zν2−1+O⁡(1ν2)),
11.11.6 𝐄ν⁡(z)=2π⁢ν⁢(sin2⁡(12⁢π⁢ν)+ν⁢zν2−1⁢cos2⁡(12⁢π⁢ν)+O⁡(1ν2)).

If ν=n(∈ℤ), then (11.10.29) applies for 𝐉n⁡(z), and

11.11.7 𝐄2⁢n⁡(z) ∼2⁢z(4⁢n2−1)⁢π,
𝐄2⁢n+1⁡(z) ∼2(2⁢n+1)⁢π,

as n→±∞.

§11.11(iii) Large ν, Fixed z/ν

For fixed λ (>0),

11.11.8 𝐀ν⁡(λ⁢ν)∼1π⁢∑k=0∞(2⁢k)!⁢ak⁡(λ)ν2⁢k+1,
ν→∞, |ph⁡ν|≤π−δ,

where

11.11.9 a0⁡(λ) =11+λ,
a1⁡(λ) =−λ2⁢(1+λ)4,
a2⁡(λ) =9⁢λ2−λ24⁢(1+λ)7,
a3⁡(λ) =−225⁢λ3−54⁢λ2+λ720⁢(1+λ)10.

In general,

11.11.9_5 ak+1⁡(λ)=λ1−λ2⁢λ⁢ak′′⁡(λ)+ak′⁡(λ)(2⁢k+1)⁢(2⁢k+2),
k=0,1,2,….

For fixed λ(>1),

11.11.10 𝐀−ν⁡(λ⁢ν)∼−1π⁢∑k=0∞(2⁢k)!⁢ak⁡(−λ)ν2⁢k+1,
ν→∞, |ph⁡ν|≤π−δ.

For fixed λ, 0<λ<1,

11.11.11 𝐀−ν⁡(λ⁢ν)∼(2π⁢ν)1/2⁢e−ν⁢μ⁢∑k=0∞(12)k⁢bk⁡(λ)νk,
ν→∞, |ph⁡ν|≤π2−δ,

where

11.11.12 μ=1−λ2−ln⁡(1+1−λ2λ),

and

11.11.13 b0⁡(λ) =1(1−λ2)1/4,
b1⁡(λ) =2+3⁢λ212⁢(1−λ2)7/4,
b2⁡(λ) =4+300⁢λ2+81⁢λ4864⁢(1−λ2)13/4.

In general,

11.11.13_5 (12)k⁢bk⁡(λ)=(−1)k(1−λ2)1/4⁢Uk⁢(11−λ2),
k=0,1,2,…,

with the Uk defined in §10.41(ii).

In particular, as ν→∞,

11.11.14 𝐀−ν⁡(λ⁢ν)∼1π⁢ν⁢(λ−1),
λ>1, |ph⁡ν|≤π−δ,
11.11.15 𝐀−ν⁡(λ⁢ν)∼(2π⁢ν)1/2⁢(1+1−λ2λ)ν⁢e−ν⁢1−λ2(1−λ2)1/4,
0<λ<1, |ph⁡ν|≤π2−δ.

Also, as ν→∞ in |ph⁡ν|≤2⁢π−δ,

11.11.16 𝐀−ν⁡(ν)∼24/337/6⁢Γ⁡(23)⁢ν1/3,

and

11.11.17 𝐀−ν⁡(ν+a⁢ν1/3)=21/3⁢ν−1/3⁢Hi⁡(−21/3⁢a)+O⁡(ν−1),

uniformly for bounded complex values of a. For the Scorer function Hi see §9.12(i).

Error bounds for (11.11.8) and (11.11.10) are given in Meijer (1932) and Nemes (2014b, c). The later references also contain exponentially-improved extensions of (11.11.8) and (11.11.10). For an extension of (11.11.17) (and (11.11.16)) into a complete asymptotic expansion, see Nemes (2020).

When ν is real and positive, all of (11.11.10)–(11.11.17) can be regarded as special cases of two asymptotic expansions given in Olver (1997b, pp. 352–360) for 𝐀−ν⁡(λ⁢ν) as ν→+∞, one being uniform for 0<λ≤1, and the other being uniform for λ≥1. (Note that Olver’s definition of 𝐀ν⁡(z) omits the factor 1/π in (11.10.4).) See also Watson (1944, §10.15).

Lastly, corresponding asymptotic approximations and expansions for 𝐉ν⁡(λ⁢ν) and 𝐄ν⁡(λ⁢ν), with 0<λ<1 or λ>1, follow from (11.10.15) and (11.10.16) and the corresponding asymptotic expansions for the Bessel functions Jν⁡(z) and Yν⁡(z); see §10.19(ii). Furthermore,

11.11.18 𝐉ν⁡(ν)∼21/332/3⁢Γ⁡(23)⁢ν1/3,
ν→∞, |ph⁡ν|≤π−δ,
11.11.19 𝐄ν⁡(ν)∼21/337/6⁢Γ⁡(23)⁢ν1/3,
ν→∞, |ph⁡ν|≤π−δ.