Page 80 - Determinants and Their Applications in Mathematical Physics
P. 80

4.3 Skew-Symmetric Determinants  65


                        =     A pr  A ps    +     A pq  A qs    +     A qr  A rs
                              A qr  A qs        A pr  A rs        A pq  A ps
                        =0.
          The theorem follows immediately if A  = 0. However, since the identity is
          purely algebraic, all the terms in the expansion of S as sums of products
          of elements must cancel out in pairs. The identity must therefore be valid
          for all values of its elements, including those values for which A = 0. The
          theorem is clearly valid if the sum is carried out over even permutations of
          any three of the four parameters.
            Notes on skew-symmetric, circulant, centrosymmetric, skew-centrosym-
          metric, persymmetric (Hankel) determinants, and symmetric Toeplitz
          determinants are given under separate headings.


          4.3 Skew-Symmetric Determinants


          4.3.1  Introduction
          The determinant A n = |a ij | n in which a ji = −a ij , which implies a ii =0,
          is said to be skew-symmetric. In detail,

                                •     a 12   a 13  a 14  ...

                                −a 12  •     a 23  a 24  ...

                                −a 13  −a 23  •  a 34  ...
                                                           .         (4.3.1)
                        A n =
                                −a 14  −a 24  −a 34  •  ...

                                .............................

                                                          n
          Theorem 4.8. The square of an arbitrary determinant of order n can be
          expressed as a symmetric determinant of order n if n is odd or a skew-
          symmetric determinant of order n if n is even.
          Proof. Let
                                     A = |a ij | n .
          Reversing the order of the rows,
                                                       n

                         A =(−1) |a n+1−i,j | n ,  N =    .          (4.3.2)
                                  N
                                                       2
          Transposing the elements of the original determinant across the secondary
          diagonal and changing the signs of the elements in the new rows 2, 4, 6,...,
                           A =(−1) |(−1) i+1 a n+1−j,n+1−i | n .     (4.3.3)
                                   N
          Hence, applying the formula for the product of two determinants in
          Section 1.4,
                          2
                         A = |a n+1−i,j | n |(−1) i+1  a n+1−j,n+1−i | n
   75   76   77   78   79   80   81   82   83   84   85