A Decision Support Tool for Smart Campus Applications – An AUS Case Study

A Master of Science thesis in Engineering Systems Management by Mohamed Faisal Khatri entitled, “A Decision Support Tool for Smart Campus Applications – An AUS Case Study”, submitted in December 2022. Thesis advisor is Dr. Vian Shawket Ahmed. Soft copy is available (Thesis, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).

Abstract

Since the last decade, advancement in the sciences around the world is causing an unstoppable incline towards smart technologies across multiple industries. Smart technologies come with numerous benefits ranging from reduced energy costs to productivity gains to lastly but very importantly, sustainability. One application of these smart technologies that attracts the objective of this thesis is “Smart campuses” which is basically the application of smart technologies to an educational institution. Few research provides the framework of the different available smart technologies that can be applied to a smart campus such as smart classrooms, smart transportation systems, smart operations and many more. However, applying all the available technologies at any campus would not be feasible in many institutes due to the various restrictions imposed by environment, culture and lack of funds, equipment, etc. Hence, this thesis involves development of a mathematical decision-making tool based on the Evidential Reasoning Approach for a successful implementation of smart applications to a university campus. The aim of this tool is to provide decision makers with rankings and utilities that enables them to decide on which smart applications (alternatives) are needed in the university based on certain inputs (attribute weights and beliefs at each evaluation grade). Upon building this tool, it was validated against 50 Truth data points from experts to analyze the quality of the tool. It was found that there is no statistically significant difference between the expert provided recommendations and the ones provided by the tool, even at a 0.05 significance level along with the area under ROC curve to be 0.734, depicting the tool performs as per expectations. Moreover, the tool’s performance evaluated using several metrics like accuracy, sensitivity etc. was upwards of 90% therefore further supporting the reliability of the tool. Lastly, the tool developed in this thesis is a generalized tool and hence can be used around the world with different number of alternatives and attributes by filling in the required inputs to the tool.