Quintic B-spline method for function reconstruction from integral values of successive subintervals (Q2249824)

Given a finite interval \([a,b]\), points \(a = x_0 < x_1 < \cdots < x_n = b\), and the values of the integrals \(\int_{x_j}^{x_{j+1}} y(x) dx\) for \(j = 0,1,\dots, n-1\), the authors develop a numerical method for approximately computing the function values \(y^{(k)}(x_j)\) for all \(j\) and \(k = 0,1,2,3,4\). The approach is based on finding a suitable quintic spline that is constructed similar to a complete interpolatory spline. Unsurprisingly, the method yields simultaneous convergence for the unknown function and its derivatives in the sense that, for \(k =0,1,2,3,4\), the \(k\)th derivative of the constructed spline approximates the \(k\)th derivative of the unknown function \(y\) in the uniform norm with an error of \(O(n^{-6+k})\).