Journal
of
Applied Mathematics and Stochastic Analysis 13:1(2000),
93-94.SttORT REPORTS AND COMMUNICATIONS
ON VACATION MODELS WITH FINITE CAPACITY
Submitted for publicationto the
Journal
of
Applied Mathematics and Stochastic AnalysisJ. LORIS-TEGHEM
University
of
Mons-Hainaut,Department of
Applied Mathematics PlaceWarocqu 1"7,
B-7000Mons,
BelgiumE-maih ][email protected]
Fax: (32)
55 37305
(Received:
December1999)
Key
words andphrases: Finite CapacityQueue,
General Vacation Policy,Queue Length.
AMS
subjectclassifications: 60K25.Contrary
to what is asserted inFrey
and Takahashi[1],
these authors were not thefirst to consider the departure epochs imbedded Markov chain for vacation models with finite capacity.
In
our paper[2],
which deals with ageneral
vacation policy, we express the stationary queuelength
distribution immediately after a departure in terms ofthecorresponding distribution in the modelwithout vacations.In [2],
we also express the stationary queuelength
distribution at an arbitrary epoch in terms ofthe corresponding distribution in the model with vacations.From
these tworelations,
one can easily derive(see [3])
the expression of the stationary queuelength
distribution at an arbitrary epoch in terms of the stationary queuelength
distribution immediately after a departure. With the notations used in[2],
thisexpression is:
L L
-1
Pv, u(J) H u(J)E(S) dL, I-I
Lu(u) (J u,...,L- 1)
pv, u(L)-
L 1-E(S) +
dLyI
L(u)"
This result contains as a particular case, the expression for the stationary queue
length
distribution at an arbitrary epoch given in[1],
for the exhaustive service multiple vacation policy.Printed in theU.S.A. (2000by North Atlantic SciencePublishing Company 93
94