A2.17

Mathematical Methods
Part II, 2001

(i) A certain physical quantity q(x)q(x) can be represented by the series n=0cnxn\sum_{n=0}^{\infty} c_{n} x^{n} in 0x<x00 \leqslant x<x_{0}, but the series diverges for x>x0x>x_{0}. Describe the Euler transformation to a new series which may enable q(x)q(x) to be computed for x>x0x>x_{0}. Give the first four terms of the new series.

Describe briefly the disadvantages of the method.

(ii) The series 1cr\sum_{1}^{\infty} c_{r} has partial sums Sn=1ncrS_{n}=\sum_{1}^{n} c_{r}. Describe Shanks' method to approximate SnS_{n} by

Sn=A+BCnS_{n}=A+B C^{n}

giving expressions for A,BA, B and CC.

Denote by BNB_{N} and CNC_{N} the values of BB and CC respectively derived from these expressions using SN1,SNS_{N-1}, S_{N} and SN+1S_{N+1} for some fixed NN. Now let A(n)A^{(n)} be the value of AA obtained from ()(*) with B=BN,C=CNB=B_{N}, C=C_{N}. Show that, if CN<1\left|C_{N}\right|<1,

1cr=limnA(n)\sum_{1}^{\infty} c_{r}=\lim _{n \rightarrow \infty} A^{(n)}

If, in fact, the partial sums satisfy

Sn=a+αcn+βdnS_{n}=a+\alpha c^{n}+\beta d^{n}

with 1>c>d1>|c|>|d|, show that

A(n)=A+γdn+o(dn)A^{(n)}=A+\gamma d^{n}+o\left(d^{n}\right)

where γ\gamma is to be found.