Page 357 - DMTH402_COMPLEX_ANALYSIS_AND_DIFFERENTIAL_GEOMETRY
P. 357

Complex Analysis and Differential Geometry




                    Notes          Let  :[a,b]  U  be a curve. A variation of  is a smooth family of curves  (t;s):[a,b] (    , ) I
                                                                                                          
                                      
                                   such that  (t;0)    (t)  for all  t [a,b].   For convenience, we will denote derivatives with respect
                                   to t as usual by a dot, and derivatives with respect to s by a prime. The generator of a variation
                                    is  the vector  field Y  (t) =  (t;  0)  along  . We  say that    is a  fixed-endpoint variation,  if
                                    (a;s)   (a),  and  (b;s)   (b)  for all  s (  , ).  Note that the generator of a fixed-endpoint
                                                  
                                                                         
                                   variation vanishes at the end points. We say that a variation  is normal if its generator Y is
                                   perpendicular to  : g   ,Y     0.  A curve  is locally L-minimizing if

                                                                       b
                                                                 L (s)   a   g     dt
                                                                            ,
                                                                  
                                   has a local minimum at s = 0 for all fixed-endpoint variations . Clearly, an L-minimizing curve
                                   is locally  L-minimizing.
                                   If  is locally L-minimizing, then any reparametrization      of  is also locally L-minimizing.

                                   Indeed, if  is  any fixed-endpoint variation of , then  (t;s)        1 (t);s   is a fixed-endpoint
                                   variation of , and since reparametrization leaves arc length invariant, we see that L (s) = L (s)
                                                                                                            
                                                                                                       
                                   which implies that L  also has a local minimum at s = 0. Thus, local minimizers of the functional
                                                   
                                   L are not necessarily parametrized proportionally to arc length. This helps clarify the following
                                   comment: a locally length-minimizing curve is not necessarily a geodesic, but according to the
                                   next theorem that is only because it may not be parametrized proportionally to arc length.
                                   Theorem 1. A locally length-minimizing curve has a geodesic reparametrization.
                                   To prove this theorem, we introduce the energy functional:
                                                                      1  b
                                                                  E    a    g    dt
                                                                            , 
                                                                     2
                                   We may now speak of  energy-minimizing and  locally energy-minimizing  curves. Our  first
                                   lemma shows the advantage of using the energy rather than the arc length functional: minimizers
                                   of E are parametrized proportionally to arc length.

                                   Lemma 1. A locally energy-minimizing curve is a geodesic.
                                   Proof. Suppose that  is a locally energy-minimizing curve. We first note that if Y is any vector
                                   field along  which vanishes at the endpoints, then setting  (t;s)    (t) sY(t),  we see that there
                                                                                           
                                   is a fixed-endpoint variation of  whose generator is Y. Since  is locally energy-minimizing, we
                                   have:

                                                                    b 1
                                                              
                                                             E (0)   a   2   g       s 0  dt  0.
                                                                          ,
                                                                              
                                                              
                                   We now observe that:
                                                                             d
                                                                  d
                                                                                j
                                                           j 
                                                                     j
                                                            s 0      dt      s 0      dt      s 0    y . j
                                                           
                                                                                   
                                                              
                                                                        
                                             i
                                   where  Y   y   is the generator of the fixed-endpoint variation , and:
                                              i
                                                                s 0   g ij,k    s 0   g ij,k  y .
                                                                         k 
                                                             g
                                                                                    k
                                                                        
                                                              ij
                                                                 
                                                                            
          350                               LOVELY PROFESSIONAL UNIVERSITY
   352   353   354   355   356   357   358   359   360   361   362