Page 68 - Essentials of applied mathematics for scientists and engineers
P. 68

book   Mobk070    March 22, 2007  11:7








                     58  ESSENTIALS OF APPLIED MATHEMATICS FOR SCIENTISTS AND ENGINEERS
                       4.2    BESSEL FUNCTIONS
                       A few differential equations are so widely useful in applied mathematics that they have been
                       named after the mathematician who first explored their theory. Such is the case with Bessel’s
                                                                           2
                       equation. It occurs in problems involving the Laplacian ∇ u in cylindrical coordinates when
                       variables are separated. Bessel’s equation is a Sturm–Liouville equation of the form

                                                    2
                                                   d u     du
                                                                         2
                                                                  2 2
                                                ρ 2    + ρ    + (λ ρ − ν )u = 0                     (4.67)
                                                   dρ 2    dρ
                       Changing the independent variable x = λρ, the equation becomes
                                                                      2
                                                                  2
                                                   x u + xu + (x − ν )u = 0                         (4.68)
                                                    2

                       4.2.1  Solutions of Bessel’s Equation
                       Recalling the standard forms (4.1) and (4.2) we see that it is a linear homogeneous equation
                       with variable coefficients and with a regular singular point at x = 0. We therefore assume a
                       solution of the form of a Frobenius series (4.17).

                                                              ∞
                                                                    j+r
                                                         u =     c j x                              (4.69)
                                                              j=0
                       Upon differentiating twice and substituting into (4.68) we find

                                   ∞
                                                                    2    j+r         j+r+2
                                      [( j + r − 1)( j + r) + ( j + r) − ν ]c j x  +  c j x  = 0    (4.70)
                                   j=0                                         j=0

                       In general ν can be any real number. We will first explore some of the properties of the solution
                       when ν is a nonnegative integer, 0, 1, 2, 3, ... . First note that

                                              ( j + r − 1)( j + r) + ( j + r) = ( j + r) 2          (4.71)

                       Shifting the exponent in the second summation and writing out the first two terms in the first


                                          (r − n)(r + n)c 0 + (r + 1 − n)(r + 1 + n)c 1 x
                                                ∞
                                                                                  j
                                             +    [(r + j − n)(r + j + n)c j + c j−2 ]x = 0         (4.72)
                                                j=2

                                                       0
                       In order for the coefficient of the x term to vanish r = n or r =−n. (This is the indicial
                       equation.) In order for the coefficient of the x term to vanish c 1 = 0. For each term in the
   63   64   65   66   67   68   69   70   71   72   73