Page 516 - Handbook Of Integral Equations
        P. 516
     where, according to (5), y(x) the inverse transform of Y(u). The right-hand side can be regarded as
               the Fourier integral of a function that is identically zero for negative x. Hence, by the uniqueness of
                                           +
               the representation of the function Y (u) by a Fourier integral, it follows that y(x) ≡ 0 on the negative
               semiaxis.
                   Conversely, if y ≡ 0 for x < 0, then the Fourier integral of this function becomes
                                                        ∞
                                                  1           iux
                                          Y(u)= √        y(x)e   dx.
                                                   2π  0
               If we replace the parameter u by a complex number z belonging to the upper half-plane, then the
               integral will converge even better. This implies the analyticity of the function
                                                   1    ∞     izx
                                           Y(z)= √        y(x)e  dx
                                                   2π  0
               in the upper half-plane.
                   The case of the lower half-plane can be treated in a similar way.
                   The integrals
                                                                        0
                                     1    ∞                       1
                             +                  izx      –                   izx
                            Y (z)= √        y(x)e  dx,  Y (z)= – √      y(x)e   dx          (7)
                                     2π  0                        2π  –∞
               are called one-sided Fourier integrals, namely, the right and the left Fourier integral, respectively.
               As well as in formula (1), the symbols ± over symbols of functions mean that the corresponding
               function is analytic in the upper or lower half-plane, respectively.
                   Let us introduce the functions
                                      y(x)  for x >0,            0     for x >0,
                              y + (x)=                   y – (x)=                           (8)
                                      0     for x <0,            –y(x)  for x <0.
               These functions are said to be one-sided functions for y(x), namely, the right function and the left
               function, respectively. Obviously, the following relation holds:
                                               y(x)= y + (x) – y – (x).                     (9)
               Applying the well-known function sign x defined by
                                                     1   for x >0,
                                            sign x =                                       (10)
                                                     –1  for x <0,
               we can express y ± in terms of y as follows:
                                                   1
                                           y ± (x)= (±1 + sign x)y(x).                     (11)
                                                   2
               The symbols ± on symbols of one-sided functions will be always subscripts.
                   The Fourier integrals of the right and left one-sided functions are the boundary values of functions
               that are analytic on the upper and lower half-planes, respectively.
                   Let us indicate the following analogs of the Sokhotski–Plemelj formulas (3) in the Fourier
               integrals:
                          1     ∞    iux
                  Y(u)= √        y(x)e  dx
                          2π  –∞
                                                    0
                          1    ∞     iux      1          iux      +      –
                      = √        y(x)e  dx + √       y(x)e  dx = Y (u) – Y (u),
                          2π  0               2π  –∞
                        ∞                                                                  (12)
                  1      Y(τ)       +     –
                              dτ = Y (u)+ Y (u)
                  πi     τ – u
                      –∞
                                ∞                   0                   ∞
                          1          iux      1          iux      1                iux
                      = √        y(x)e  dx – √       y(x)e  dx = √       y(x) sign xe  dx.
                          2π  0               2π  –∞              2π  –∞
                 © 1998 by CRC Press LLC
               © 1998 by CRC Press LLC
                                                                                                             Page 498
     	
