Page 195 - Determinants and Their Applications in Mathematical Physics
P. 195

180   5. Further Determinant Theory

            Assume that
                               Pf m = H m K m ,  m < n.             (5.2.10)
          The method by which the theorem is proved for all values of n is outlined
          as follows.
            Pf n is expressible in terms of Pfaffians of lower order by the formula
                                   2n−1
                                           i+1  (n)
                             Pf n =    (−1)   Pf  a i,2n ,          (5.2.11)
                                                i
                                    i=1
          where, in this context, a i,2n is defined as a sum in (5.2.4) so that Pf n
          is expressible as a double sum. The introduction of a variable x enables
          the inductive assumption (5.2.10) to be expressed as the equality of two
          polynomials in x. By equating coefficients of one particular power of x,an
                                         (n)
          identity is found which expresses Pf  as the sum of products of cofactors
                                         i
          of H n and K n (Lemma 5.5). Hence, Pf n is expressible as a triple sum
          containing the cofactors of H n and K n . Finally, with the aid of an identity
          in Appendix A.3, it is shown that the triple sum simplifies to the product
          H n K n .
            The following Pfaffian identities will also be applied.

                                        (n)     (2n−1) 1/2

                                      Pf   = A          ,           (5.2.12)
                                        i       ii
                                    (n)  (n)   (2n−1)
                           (−1) i+j Pf  Pf  = A     ,               (5.2.13)
                                    i   j      ij
                                       (n)
                                    Pf     =Pf n−1 .                (5.2.14)
                                       2n−1
          The proof proceeds with a series of lemmas.
          5.2.2  Four Lemmas
          Let a ∗  be the function obtained from a ij by replacing each φ r by
               ij
          (φ r − xφ r+1 ) and by replacing each ψ r by (ψ r − xψ r+1 ).
          Lemma 5.2.
                                                           2
                       a = a ij − (a i,j+1 + a i+1,j )x + a i+1,j+1 x .
                        ∗
                        ij
          Proof.
                            j−i

                       a =     (φ s+i−1 − xφ s+i )(ψ j−s − xψ j−s+1 )
                        ∗
                        ij
                            s=1
                                               2
                          = a ij − (s 1 + s 2 )x + s 3 x ,
          where
                                     j−i

                                s 1 =   φ s+i−1 ψ j−s+1
                                     s=1
                                  = a i,j+1 − φ i ψ j ,
   190   191   192   193   194   195   196   197   198   199   200