Keywords (Extracted from title, table of contents and
abstract of thesis) Analysis, Network, Queues,
Finite, Capacities, Blocking, numerical solutions, analytic,
algorithms 
Abstract Queueing Network
Models (QNMs) with Finite Capacity provide powerful and realistic
tools for the performance evaluation and prediction of discrete flow
systems such as computer systems, communication networks and
flexible manufacturing systems. Over recent years, there has been a
great deal of progress towards the analysis and application of QNMs
with finite capacity, and high quality research work has appeared in
diverse scientific journals of learning and conference proceedings
in the fields of Operations Research, Computer Science,
Telecommunication Networks, Management and Industrial Engineering.
However, there are still many important and interesting finite
capacity queues and QNMs to be resolved, such as those involving
multiplejob classes, bounds and theoretical properties, exact
analysis, numerical solutions and approximate methods, as well as
application studies to computer and distributed systems, highspeed
networks and production systems.
Finite capacity queueing network models (QNMs) also play an
important role towards effective congestion control and quality of
service (QoS) protection of modern discrete flow networks. Blocking
in such networks arises because the traffic of jobs through one
queue may be momentarily halted if the destination queue has reached
its capacity. Exact closedform solutions for QNMs with blocking are
not generally attainable except for some special cases such as
twostation cyclic queues and ‘reversible’ queueing networks. As a
consequence, numerical techniques and analytic approximations have
been proposed for the study of arbitrary QNMs with nonMarkovian
(external) interarrival and service times under various types of
blocking mechanisms. This research mainly focuses on:
i) To develop and validate cost effective analytical models for
arbitrary QNMs with blocking and multiple job classes.
ii) To use the analytical models to evaluate the performance of QNMs
under various blocking mechanisms applicable to flexible
manufacturing systems and high speed telecommunication networks.
iii) To develop approximate analytical algorithms for arbitrary QNMs
consisting of G/G/1/N censoredtype queues with arbitrary
arrival and service processes, single server under Partial Buffer
Sharing (PBS) and Complete Buffer Partitioning (CBP) schemes
stipulating a sequence of buffer thresholds {N=N1,N2,…,NR,0< Ni ≤
Ni1 , i=1,2,…,R} and buffer partitioning with FCFS service
discipline. {chapter 4 and 5}
iv) Validation of these algorithms (iii) using QNAP simulation
package.
v) Extension of the above algorithms for multiple servers and its
validation using simulation.
Determining a performance distribution via classical queueing theory
may prove to be an infeasible task even for systems of queues with
moderate complexity. Hence, the principle of entropy maximization
may
be applied to characterize useful information theoretic
approximations
of performance distributions of queueing systems and queueing
network
models (QNMs).
Focusing on an arbitrary open QNM, the ME solution for the joint
state
probability, subject to marginal mean value constraints, can be
interpreted as a productform approximation. Thus, the principle of
ME
implies a decomposition of a complex QNM into individual queues each
of which can be analyzed separately with revised inter arrival and
service times. Moreover, the marginal ME state probability of a
single
queue, in conjunction with suitable formulae for the first two
moments
of the effective flow, can play the role of a costeffective
analytic
building block towards the computation of the performance metrics
for
the entire network.
