Evaluation of the Performance of Customer Service Representatives in a Call Center Using DEA/ Network Model/ Fuzzy Sets

• Main Author: Poykayil Jayananda Panicker, Abhishek
• Other Authors: Industrial and Systems Engineering
• Format: Others
• Published: Virginia Tech 2014
• Subjects: Fuzzy; DEA; Performance; Productivity; Efficiency
• Online Access: http://hdl.handle.net/10919/31704; http://scholar.lib.vt.edu/theses/available/etd-04132003-184221/

Data Envelopment Analysis (DEA) is a linear programming technique that has been used extensively in the literature to measure relative efficiency. One of the main attributes of DEA is that it can model multiple inputs and multiple outputs in the model. In this research work, attributes pertaining to service quality have been modeled using the Network model.The primary research is the augmentation of the existing Network model to include input/ output variables that are imprecise from a measurement point of view. These variables are qualitative assessments that have a linguistic representation/ interpretation. A very good example of this variable would be "Pleasantness". Given the fact that there are different evaluators, there is a certain degree of impreciseness associated with the representation of each of these qualitative variables. This imprecision is captured using the fuzzy sets. The triangular membership functions were used to describe the membership functions. So a unique network model that captured fuzzy variables was created. The second main research contribution is that this is the first attempt of capturing service quality and efficiency of customer service representatives. The generic model that was created was used to evaluate the performance of the customer service representatives in a major airline. The results that were obtained, was shared with the decision makers at the airline for validation. The results that were obtained from the model also helped us validate the model with the other existing models.One of the main advantages of using the DEA/ Network/ Fuzzy model was that the imprecision involved in measuring the customer service representatives were accounted for. This enabled the decision maker in making the right decisions and not penalizing a customer service representative for imprecision in the data. Graphical Interpretations were also provided for the results that were obtained from the analysis. === Master of Science