Reliability Evaluation and Topological Design of Communication Networks

Abstract

newline The performance of any communication network depends upon the probability of newlinesuccessful data transmission or in other words performance of the network is defined as newlinethe communicability of the nodes. It depends upon many factors in which reliability play newlinean important role. The reliable communication between the users is the prime concern of newlinethe modern communication scenario. However, the exact reliability evaluation of any newlinecommunication networks is important in order to find optimal topology design polices, newlineparticularly when there exist imperfect hop coverage in between two nodes, cost of the newlinelayout design, common-cause failures, and delay etc. Therefore, in this thesis two aspects newlineof the complex networks, namely, (i) evaluation of reliability indices, and (ii) optimal newlinetopological design of networks in order to maximize reliability and minimize the cost and newlinedelay are studied. newlineIn networks, the performance depends upon the probability of a specified set of nodes newlinebeing communicable. Among the several approaches of reliability evaluation, the newlinemultiple-variable-inversion sum-of-disjoint product (MVI-SDP) approach provides the newlinereliability expression in a most efficient and compact manner. However, it needs newlineefficiently enumerated inputs depending on the output requirement i.e. required reliability newlinemeasures (as input will be different for each reliability measure). Therefore, the major newlinecontribution in this work is to provide a minimal pathset or cutset based common newlineframework to evaluate network reliability measures viz., 2-, g-, and k-terminal reliability. newlineThe work proposes various efficient algorithms to enumerate spanning trees and gminimal newlinecutsets, which is a new idea, directly from the adjacency matrix representation of newlinea network. These spanning trees or g-minimal cutsets can be used in conjunction with newlineMVI-SDP approach to evaluate g-terminal reliability or global reliability. Another newlineattempt has been made in this objective to provide an answer to the question as to when newlinecutsets or pathsets are suitable for evaluating reliability measures. An exhaustive study newlinehas been conducted to provide some guidelines in this respect.