A note on the radius of convergence of the USA algorithm

From MaRDI portal





The authors described previously an algorithm termed the method of updated successive approximation, or USA for short [ibid. 9, No.2, 127- 137 (1985; Zbl 0663.90011)]. USA is a fixed-point algorithm that converges to the true equilibrium for an arbitrary choice of initial conditions. This paper extends the results developed in the original paper by showing that the algorithm has an arbitrarily large radius of convergence. The radius of convergence is used to address an issue raised by \textit{D. Herceg} and \textit{L. Cvetković} [ibid. 13, No.2, 301-311 (1989; Zbl 0663.90012)]. Herceg and Cvetkovic present an example which they claim shows the nonconvergence of the USA algorithm for a particular set of initial conditions. A recomputation of their example with identical initial conditions but arbitrarily large radius of convergence shows that the algorithm does indeed converge to the true equilibrium as predicted.











This page was built for publication: A note on the radius of convergence of the USA algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1115333)