A comparative analysis of the successive lumping and the lattice path counting algorithms (Q2804417)

The paper continues the investigations of the same authors and compares the successive lumping (SL) methodology with the lattice path counting algorithm (LPCA, see [\textit{J. S. H. van Leeuwaarden} and \textit{E. M. M. Winands}, Stoch. Models 22, No. 1, 77--98 (2006; Zbl 1115.60070)] and [\textit{J. S. H. van Leeuwaarden} et al., J. Appl. Probab. 46, No. 2, 507--520 (2009; Zbl 1186.60087)]) from the viewpoints of applicability requirements and numerical complexity for models where the objective is the calculation of the steady-state distribution of pertinent quasi birth-and-death processes, i.\,e., two-dimensional Markov chains allowing only transitions to the left and to the right in every state. It is shown that SL often yields faster algorithms than LPCA, and the fields of applicability partly overlap. The paper continues to specialize the results to homogeneous QBD in order to make a comparison.