Page 145 - Determinants and Their Applications in Mathematical Physics
P. 145

130   4. Particular Determinants

          Referring to the section on differences in Appendix A.8,

                                            m
                                     φ m =∆ θ 0
          so that
                                       B = A.
          The Hankelian B arises in studies by M. Yamazaki and Hori of the Ernst
          equation of general relativity and A arises in a related paper by Vein.
            Define determinants U(x), V (x), and W, each of order (n + 1), by bor-
          dering A in different ways. Since a ij is a function of x and y, it follows that
          U(x) and V (x) are also functions of y. The argument x in U(x) and V (x)
          refers to the variable which appears explicitly in the last row or column.
                                                   x


                                                  x /3
                                                   3

                                                  x /5
                                                   5
                                    [a ij ] n
                   U(x)=

                                                   ···

                                             x
                                              2n−1
                                                  /(2n − 1)
                           111       ···  1        •

                                                           n+1
                            n   n      2r−1
                                   A rs x
                        = −                ,                       (4.10.24)
                                    2r − 1
                            r=1 s=1
                                                        1


                                                       1/3


                                                       1/5

                                      [a ij ] n
                   V (x)=
                                                       ···


                                                    1/(2n − 1)
                           xx     x    ···  x           •
                               3   5          2n−1
                                                              n+1
                            n   n      2s−1
                                   A rs x
                        = −                ,                       (4.10.25)
                                    2r − 1
                            r=1 s=1
                     W = U(1) = V (1).                             (4.10.26)
          Theorem 4.39.
                                      2
                                                 2
                                         2
                             2
                               2
                            p U (x)+ q U (y)= W − AW.
          Proof.
                                      A is x  2i−1    A jr x 2j−1
                                               n
                            2

                          U (x)=
                                       2i − 1      2j − 1
                                   i,s        j,r
                                       A is A jr x 2(i+j−1)
                                =                    .
                                       (2i − 1)(2j − 1)
                                  i,j,r,s
          Hence,
                    2
                 2
                             2
                          2
                p U (x)+ q U (y) − W 2
   140   141   142   143   144   145   146   147   148   149   150