Answers are grounded in this paper and cite their sources.
For Markov renewal processes of GI/M/1 type, we define a double transform R(z,s), a generalization of Neuts’ rate matrix R, and establish an interesting duality theorem. That result expresses R(z,s) explicitly in terms of the double transform G(z,s) of Neuts for a suitably defined dual process of M/G/1 type. For the GI/PH/1 queue, the G(z,s) matrix arising in the duality theorem is identified as the matrix of the dual PH/G/1 queue with a time reversed representation of the PH-distribution. The duality theorem unifies the results on the two paradigms and enables one to obtain many results concerning one from those of the other by simple arguments.
No takes yet. Share an insight, caveat, or question.
V. Ramaswami (1990) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: