Optimal recovery of linear functionals on sets of finite dimension (Q2518078)

Let \(X\) be a linear space of dimension \(n+1\) generated by \(f_0, f_1,\dots,f_n\). Let \(L_1,\dots,L_n\) be linear functionals linearly independent defined on \(X\) and \(L_0\neq L_i\), \(i=1,\dots,n\), another linear functional on \(X\). Let \(P=\{\sum^n_{i=0}a_if_i:a_i\in\mathbb R\), \(|a_i|\leq |\beta_i|\), \(i=0,\dots,n\}\), where \(\beta n=1\) and \((\beta_0,\dots,\beta n_1)\) is the solution of the system \(\sum^n_{i=0} \beta j\) \(L_jf_i=0\), \(j=1,\dots,n\). The author presents an algorithm that recovers the functional \(L_0\) on \(P\) with the least error among all linear algorithms using the information \((L_1f,\dots,L_nf)\).