Solution of integral equations using generalised inverse, function-valued Padé approximants. I (Q751224)

Rational functions of \(\lambda\) are derived from the series \(f(x,\lambda)=\sum c(j,x)\lambda^ j\) (0\(\leq j\leq \infty)\) where c(j)\(\in {\mathcal L}_ 2(a,b)\). With n,k\(\geq 0\) fixed, the integrals \(M(i,j)=\sum \int^{b}_{a}c(\tau +i+n-2k+1,x)c(j-\tau +n-2k,x)^*dx\) (0\(\leq \tau \leq j-i-1)\) with \(0\leq i\leq 2k\) and \(i<j\leq 2k\) are evaluated. q(n,2k;\(\lambda\)) is the value of the \((2k+1)th\) order determinant whose ith row (1\(\leq i\leq 2k)\) is -M(j-1,i-1) \((1\leq j<i-1)\), 0 and M(i-1,j) (i\(\leq j\leq 2k)\) and whose \((2k+1)th\) row is \(\lambda^{2k-j+1}\) \((1\leq j\leq 2k+1)\). With \(q(n,2k;\lambda)f(x,\lambda)=\sum r(j,x)\lambda^ j\) (0\(\leq j\leq \infty)\), p(n,2k;x,\(\lambda\)) is \(\sum r(j,x)\lambda^ j\) (0\(\leq j\leq n)\). p(n,2k;x,\(\lambda\))/q(n,2k;\(\lambda\)) is the (n,2k) approximant to f(x,\(\lambda\)). Series having the form of f(x,\(\lambda\)) arise in the solution of integral equations and it is in this context that the use of the above approximants is illustrated.