Page 194 - Schaum's Outline of Theory and Problems of Advanced Calculus
P. 194

CHAP. 8]               APPLICATIONS OF PARTIAL DERIVATIVES                      185


                                                                                         dR
                                                                                              .If r 0 and r

                                                                                          du    P
                        If R ¼ f ðuÞi þ gðuÞj þ hðuÞk,a vector tangent to C at the point P is given by T 0 ¼
                     denote the vectors drawn respectively from O to Pðx 0 ; y 0 ; z 0 Þ and Qðx; y; zÞ on the tangent line, then since
                     r   r 0 is collinear with T 0 we have

                                                                     dR
                                                                          ¼ 0
                                                ðr   r 0 Þ  T 0 ¼ðr   r 0 Þ                          ð5Þ
                                                                     du
                                                                        P
                     In rectangular form this becomes
                                                     x   x 0  y   y 0  z   z 0
                                                          ¼      ¼                                   ð6Þ
                                                      0      0       0
                                                     f ðu 0 Þ  g ðu 0 Þ  h ðu 0 Þ
                     The parametric form is obtained by setting each ratio equal to u.
                        If the curve C is given as the intersection of two surfaces with equations Fðx; y; zÞ¼ 0 and
                     Gðx; y; zÞ¼ 0 observe that rF  rG has the direction of the line of intersection of the tangent planes;
                     therefore, the corresponding equations of the tangent line are
                                                x   x 0     y   y 0    z   z 0
                                                        ¼          ¼
                                                                                                     ð7Þ

                                                F y        F z        F x
                                                    F z        F x        F y

                                                G y  G z P  G z  G x P  G x  G y P
                     Note that the determinants in (7) are Jacobians.  A similar result can be found when the surfaces are
                     given in terms of orthogonal curvilinear coordinates.
                     4.  Normal Plane to a Curve.  Suppose we wish to find an equation for the normal plane to curve C
                     at Pðx 0 ; y 0 ; z 0 Þ of Fig. 8-2 (i.e., the plane perpendicular to the tangent line to C at this point). Letting r be
                     the vector from O to any point ðx; y; zÞ on this plane, it follows that r   r 0 is perpendicular to T 0 . Then
                     the required equation is


                                                                    dR
                                                                        ¼ 0

                                                 ðr   r 0 Þ  T 0 ¼ðr   r 0 Þ                         ð8Þ
                                                                    du    P
                        When the curve has parametric equations x ¼ f ðuÞ; y ¼ gðuÞ; z ¼ hðuÞ this becomes
                                          f ðu 0 Þðx   x 0 Þþ g ðu 0 Þð y   y 0 Þþ h ðu 0 Þðz   z 0 Þ¼ 0  ð9Þ
                                                                      0
                                           0
                                                        0
                        Furthermore, when the curve is the intersection of the implicitly defined surfaces
                                                Fðx; y; zÞ¼ 0and Gðx; y; zÞ¼ 0
                     then

                                       F y              F z               F x
                                         F z              F x               F y    ðz   z 0 Þ¼ 0
                                             ðx   x 0 Þþ      ð y   y 0 Þþ                          ð10Þ

                                     G y  G z P       G z  G x P       G x  G y P
                     5.  Envelopes.  Solutions of differential equations in two variables are geometrically represented by
                     one-parameter families of curves.  Sometimes such a family characterizes a curve called an envelope.
                        For example, the family of all lines (see Problem 8.9) one unit from the origin may be represented by
                                                                                               2
                                                                                                   2
                     x sin     y cos     1 ¼ 0, where   is a parameter. The envelope of this family is the circle x þ y ¼ 1.
                        If  ðx; y; zÞ¼ 0isa one-parameter family of curves in the xy plane, there may be a curve E which is
                     tangent at each point to some member of the family and such that each member of the family is tangent
                     to E.If E exists, its equation can be found by solving simultaneously the equations
                                                  ðx; y; Þ¼ 0;      ðx; y; Þ¼ 0                     ð11Þ

                     and E is called the envelope of the family.
   189   190   191   192   193   194   195   196   197   198   199