Analytic extrapolation in \(L^{\infty}\)-norm: An alternative approach to ``QCD sum rules'' (Q1114908)

In view of physical applications (especially to ``QCD Sum Rules''), the following problem, pertaining to analytic extrapolation techniques, is studied. We are considering ``amplitudes'', which are (real) analytic functions in the complex plane cut along \(\Gamma =[s_ 0,\infty)\). A model \(F_ 0(s)\) of the amplitude is given through the values of \(F_ 0(s)\) on some interval \(\gamma =[s_ 2,s_ 1]\) (with \(s_ 1<s_ 0)\) and the values of its discontinuity on \(\Gamma\). These values are approximate, and are supplemented by prescribed error channels, measured in \(L^{\infty}\)-norm (both on \(\Gamma\) and \(\gamma)\). Investigating the compatibility between these data leads to an extremum problem which is solved up to a point where numerical methods can be implemented.