Page 103 - Determinants and Their Applications in Mathematical Physics
P. 103

88   4. Particular Determinants

          which is centrosymmetric and can therefore be expressed as the prod-
          uct of two determinants of lower order. T n is also persymmetric about
          its secondary diagonal.
            Let A n , B n , and E n denote Hankel matrices defined as follows:

                                                ,
                                   A n = t i+j−2
                                               n

                                                ,
                                   B n = t i+j−1
                                               n

                                              .                      (4.5.8)
                                   E n = t i+j
                                             n
          Then, the factors of T n can be expressed as follows:
                                  1
                          T 2n−1 = |T n−1 − E n−1 ||T n + A n |,
                                  2
                            T 2n = |T n + B n ||T n − B n |.         (4.5.9)
          Let
                                           1
                              1
                         P n = |T n − E n | = |t |i−j| − t i+j | n ,
                              2            2
                                           1
                              1
                         Q n = |T n + A n | = |t |i−j| + t i+j−2 | n ,
                              2            2
                                           1
                              1
                         R n = |T n + B n | = |t |i−j| + t i+j−1 | n ,
                              2            2
                                           1
                              1
                         S n = |T n − B n | = |t |i−j| − t i+j−1 | n ,  (4.5.10)
                              2            2
                         U n = R n + S n ,
                         V n = R n − S n .                          (4.5.11)
          Then,
                                  T 2n−1 =2P n−1 Q n ,
                                    T 2n =4R n S n
                                           2
                                                2
                                       = U − V .                    (4.5.12)
                                           n   n
          Theorem.
          a. T 2n−1 = U n−1 U n − V n−1 V n ,
          b. T 2n = P n Q n + P n−1 Q n+1 .
          Proof. Applying the Jacobi identity (Section 3.6),
                                  (n)  (n)
                                T     T
                                 11
                                                 (n)
                                       1n    = T n T  .
                                 (n)
                                T n1  T nn
                                       (n)      1n,1n
          But
                                   (n)    (n)
                                  T   = T   = T n−1 ,
                                   11    nn
                                   (n)    (n)
                                  T   = T   ,
                                   n1    1n
                                  (n)
                                T     = T n−2 .
                                 1n,1n
          Hence,


                              T 2  = T n T n−2 + T  (n) 2 .         (4.5.13)
                                n−1             1n
   98   99   100   101   102   103   104   105   106   107   108