University of Kent,
Introducing Combined Weights and Centrality Measures To Evaluate Network Topologies
Abstract
dc:description.abstractIn this research of network structure analysis, the knowledge of centrality measures is applied to discover or predict a most important actor or node in a network/graph. Problems of energy eciency and sustainability are considered, and also those of allocation of resources. In order to enable an ecient allocation of energy resources to the right path in a distributed network such as obtained in a data center, author ́s network and supply chain network, new measures of centralities are introduced aside from the traditional ones of Closeness, Betweenness, Degree and Eigen-Vector centralities. Mixed-mean centrality, which is based on the generalized degree centrality, was developed as a measure to emphasise the importance of a node in the authorship network and the distributed system of a data centre. Weighted centrality measure when used as against the traditional measures mentioned above was able to make prediction for a Distribution Centre (DC) of a Supply Chain Network with an accuracy of 91.6%. Clique-Structure/Node-weighted centrality measure was able to make a prediction with 66% accuracy, while the Weighted Marking, Clique-Structure/Node-Weighted Centrality made a prediction accuracy of 96.2%. The Top Eigen-Vector Weighted Centrality and Newtonian Gravitational Force were also used to predict the location of distribution centre (DC) in a supply chain network with accuracies of 92.9% and 96.9% respectively.
Degree
thesis:*- Name dc:type.qualificationname
- phd
- Level dc:type.qualificationlevel
- doctoral
- Grantor dc:publisher.institution
- University of Kent,
- Year dc:date.issued
- 2017
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Akanmu, Amidu Akinpelumi Gbolasere
- Advisor dc:contributor.advisor
-
- Wang, Frank Z.
Rights
- Language dc:language
- en