Data-driven Service Operations Management
Public DepositedAdd to collection
You do not have access to any existing collections. You may create a new collection.
Downloadable Content
Download PDFCitation
MLA
Ye, Han. Data-driven Service Operations Management. University of North Carolina at Chapel Hill, 2014. https://doi.org/10.17615/z7yf-jm82APA
Ye, H. (2014). Data-driven Service Operations Management. University of North Carolina at Chapel Hill. https://doi.org/10.17615/z7yf-jm82Chicago
Ye, Han. 2014. Data-Driven Service Operations Management. University of North Carolina at Chapel Hill. https://doi.org/10.17615/z7yf-jm82- Last Modified
- March 19, 2019
- Creator
-
Ye, Han
- Affiliation: College of Arts and Sciences, Department of Statistics and Operations Research
- Abstract
- This dissertation concerns data driven service operations management and includes three projects. An important aim of this work is to integrate the use of rigorous and robust statistical methods into the development and analysis of service operations management problems. We develop methods that take into account demand arrival rate uncertainty and workforce operational heterogeneity. We consider the particular application of call centers, which have become a major communication channel between modern commerce and its customers. The developed tools and lessons learned have general appeal to other labor-intensive services such as healthcare. The first project concerns forecasting and scheduling with a single uncertain arrival customer stream, which can be handled by parametric stochastic programming models. Theoretical properties of parametric stochastic programming models with and without recourse actions are proved, that optimal solutions to the relaxed programs are stable under perturbations of the stochastic model parameters. We prove that the parametric stochastic programming approach meets the quality of service constraints and minimizes staffing costs in the long-run. The second project considers forecasting and staffing call centers with multiple interdependent uncertain arrival streams. We first develop general statistical models that can simultaneously forecast multiple-stream arrival rates that exhibit inter-stream dependence. The models take into account several types of inter-stream dependence. With distributional forecasts, we then implement a chance-constraint staffing algorithm to generate staffing vectors and further assess the operational effects of incorporating such inter-stream dependence, considering several system designs. Experiments using real call center data demonstrate practical applicability of our proposed approach under different staffing designs. An extensive set of simulations is performed to further investigate how the forecasting and operational benefits of the multiple-stream approach vary by the type, direction, and strength of inter-stream dependence, as well as system design. Managerial insights are discussed regarding how and when to take operational advantage of the inter-stream dependence. The third project of this dissertation studies operational heterogeneity of call center agents with regard to service efficiency and service quality. The proxies considered for agent service efficiency and service quality are agents' service times and issue resolution probabilities, respectively. Detailed analysis of agents' learning curves of service times are provided. We develop a new method to rank agents' first call resolution probabilities based on customer call-back rates. The ranking accuracy is studied and the comparison with traditional survey-driven methods is discussed.
- Date of publication
- 2014
- Keyword
- DOI
- Resource type
- Rights statement
- In Copyright
- Advisor
- Shen, Haipeng
- Degree
- Doctor of Philosophy
- Graduation year
- 2014
- Language
- Publisher
Relations
- Parents:
This work has no parents.
Items
Thumbnail | Title | Date Uploaded | Visibility | Actions |
---|---|---|---|---|
Ye_unc_0153D_14302.pdf | 2019-04-11 | Public | Download |