Numerical computation of the half Laplacian by means of a fast convolution algorithm

Authors C. M. Cuesta, Francisco de la Hoz, Ivan Girona

Classical hypergeometric functions, ({}_2F_1) (33C05) Fractional derivatives and integrals (26A33) Numerical methods for discrete and fast Fourier transforms (65T50)

Abstract: In this paper, we develop a fast and accurate pseudospectral method to approximate numerically the fractional Laplacian

(- D e l t a)^{a l p h a / 2}

of a function on

m a t h b b R

for

a l p h a = 1

; this case, commonly referred to as the half Laplacian, is equivalent to the Hilbert transform of the derivative of the function. The main ideas are as follows. Given a twice continuously differentiable bounded function

u i n m a t h i t C_{b}^{2} (m a t h b b R)

, we apply the change of variable

x = L c o t (s)

, with

L > 0

and

s i n [0, p i]

, which maps

m a t h b b R

into

[0, p i]

, and denote

(- D e l t a)_{s}^{1 / 2} u (x (s)) e q u i v (- D e l t a)^{1 / 2} u (x)

. Therefore, by performing a Fourier series expansion of

u (x (s))

, the problem is reduced to computing

(- D e l t a)_{s}^{1 / 2} e^{i k s} e q u i v (- D e l t a)^{1 / 2} (x + i)^{k} / (1 + x^{2})^{k / 2}

. On a previous work, we considered the case with

k

even and

a l p h a i n (0, 2)

, so we focus now on the case with

k

odd. More precisely, we express

(- D e l t a)_{s}^{1 / 2} e^{i k s}

for

k

odd in terms of the Gaussian hypergeometric function

_{2} F_{1}

, and also as a well-conditioned finite sum. Then, we use a fast convolution result, that enable us to compute very efficiently

s u m_{l = 0}^{M} a_{l} (- D e l t a)_{s}^{1 / 2} e^{i (2 l + 1) s}

, for extremely large values of

M

. This enables us to approximate

(- D e l t a)_{s}^{1 / 2} u (x (s))

in a fast and accurate way, especially when

u (x (s))

is not periodic of period

p i

.

This page was built for publication: Numerical computation of the half Laplacian by means of a fast convolution algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6510481)