Answers are grounded in this paper and cite their sources.
For an irreducible Markov-modulated Poisson process, let T^π be the time from a stationary instant to the next arrival and T^α the time from an arrival to the next one in the event-stationary version. Asanjarani and Nazarathy conjectured that T^π is stochastically larger than T^α, and He listed the same property among open questions on the MMPP class of phase-type distributions. The conjecture is false. The order is equivalent to the hazard rate of T^π never exceeding the arrival rate. For a Cox process driven by a non-constant periodic intensity with uniform phase this fails on a set of positive measure. For the MMPP on the directed n-cycle with arrival rates ε(1+bcos(2π j/n)) it fails for all small ε>0 whenever n≥ 12, and an explicit twelve-state MMPP with positive rational rates has P(T^π>19/2)- P(T^α>19/2)<-1/300000, proved in exact rational arithmetic. For reversible modulators the order holds. We also show that the mean of He's MMPP-type distribution need not be convex in the speed of modulation.
No takes yet. Share an insight, caveat, or question.
Yitzchak Shmalo (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: