Advisory Committee Chair
Murat M Tanik
Advisory Committee Members
Gary J Grimes
Chittoor V Ramamoorthy
Murat M Tanju
Gregg L Vaughn
B Earl Wells
Document Type
Dissertation
Date of Award
2007
Degree Name by School
Doctor of Philosophy (PhD) School of Engineering
Abstract
Component-based software systems are becoming larger and more complex. The quantification of these systems is possible through representation in a hierarchical structure with the use of software metrics. Metrics provide information about such features as complexity, design, quality, size, development time, effort, and cost. Entropy-based software metrics promise important improvements in measuring complexity, design quality, and information flow. The entropy-based measurement framework, the core of this dissertation, offers quantitative representation and decomposition processes to apply the entropy-based metrics that we developed: Design Information Content, Interaction Complexity, Interaction Coupling, and Interaction Cohesion. The objective is to achieve a hierarchical representation of the system in which we can see the logic flow and quantify the interaction among the components. The framework is useful to both the designer and the user in the design and maintenance phases to detect structural complexity issues and comprehend the interaction of the software modules. Another motivation of this work is to lead the way to the development of measurement software which uses entropy-based methods. The representationdecomposition methods are based on graph theory, mainly cubic control flowgraphs, and the computation methods are based on information theory, mainly entropy; thus, the entropy- based measurement framework has a solid foundation to become an analytical measurement framework for component-based systems.
Recommended Citation
Aktunc, Ozgur, "An Entropy-Based Measurement Framework For Component-Based Hierarchical Systems" (2007). All ETDs from UAB. 3654.
https://digitalcommons.library.uab.edu/etd-collection/3654