Page 49 - Essentials of applied mathematics for scientists and engineers
P. 49

book   Mobk070    March 22, 2007  11:7








                                                                 ORTHOGONAL SETS OF FUNCTIONS        39
                   3.3    STURM–LIOUVILLE PROBLEMS: ORTHOGONAL
                          FUNCTIONS
                   We now proceed to show that solutions of a certain ordinary differential equation with certain
                   boundary conditions (called a Sturm–Liouville problem) are orthogonal functions with respect to a
                   weighting function, and that therefore a well-behaved function can be represented by an infinite
                   series of these orthogonal functions (called eigenfunctions), as in Eqs. (3.12) and (3.16).
                        Recall that the problem
                                            2
                                     X xx + λ X = 0, X(0) = 0, X(1) = 0   0 ≤ x ≤ 1             (3.40)
                   has solutions only for λ = nπ and that the solutions, sin(nπx) are orthogonal on the interval
                   (0, 1). The sine functions are called eigenfunctions and λ = nπ are called eigenvalues.
                        As another example, consider the problem

                                                             2
                                                     X xx + λ X = 0                             (3.41)
                   with boundary conditions

                                                   X(0) = 0
                                                                                                (3.42)
                                                   X(1) + HX x (1) = 0
                   The solution of the differential equation is

                                                X = A sin(λx) + B cos (λx))                     (3.43)

                   The first boundary condition guarantees that B = 0. The second boundary condition is satisfied
                   by the equation

                                                A[sin(λ) + Hλ cos(λ)] = 0                       (3.44)
                   Since A cannot be zero, this implies that

                                                     − tan(λ) = Hλ.                             (3.45)

                   The eigenfunctions are sin(λx)and the eigenvalues are solutions of Eq. (3.45). This is illustrated
                   graphically in Fig. 3.4.
                        We will generally be interested in the fairly general linear second-order differential
                   equation and boundary conditions given in Eqs. (3.46) and (3.47).
                                       d      dX
                                          r(x)     + [q(x) + λp(x)]X = 0 a ≤ x ≤ b              (3.46)
                                      dx      dx
                                      a 1 X(a) + a 2 dX(a)/dx = 0

                                      b 1 X(b) + b 2 dX(b)/dx = 0                               (3.47)
   44   45   46   47   48   49   50   51   52   53   54