Service Engineering: Data-Based Course Development and Teaching

Abstract

I n this exposition, we discuss empirically based teaching in the newly emerging field of "service engineering." Specifically, we survey a service engineering course, taught at the Technion-Israel Institute of Technology. The course was "born" about 15 years ago as a graduate seminar and ultimately took its present form as a core course for the undergraduate program in industrial engineering and management. The role of measurements and data as teaching-enhancers and research-drivers is underscored. In addition, we emphasize that data granularity must reach the individual-transaction level. We describe customized databases and software tools that facilitate operational and statistical analysis of services; this includes the use of SEEStat, a data-user interface that was developed at the Technion's SEE Laboratory, for research and educational purposes. Some unique aspects of the course are the incorporation of state-of-the-art research and real-world data in lectures, recitations, and home assignments, as amply presented throughout this work. The application focus of the surveyed course has been telephone call centers, which constitute an explosively growing branch of the service industry. The course is now expanding to also cover healthcare, especially hospitals; some examples from other service areas (e.g., the justice system) are described as well.

Service Engineering: Data-Based Course Development and Teaching

“…Erlang's models (Erlang 1948) are standardly taught in elementary/introductory academic courses (for example, M/M/n, M/M/n/n), as they are still cornerstones of today's telecommunication models (where M/M/n/n is known as Erlang-B, "B" apparently for blocking-the central feature of this modeland M/M/n is referred to as Erlang-C, "C" conceivably because it is subsequent to "B"). Moreover, and more relevant to our present discussion, M/M/n is still the workhorse that supports workforce decisions in many telephone call centers; see §4.6 of Full Version (Mandelbaum and Zeltyn 2009).…”

Inter-dependent, heterogeneous, and time-varying service-time distributions in call centers

“…Motivated by this, Shen and Brown (2006) proposed a new method for inference about non-parametric regression curves when the errors are lognormally distributed, and illustrated their method with both a simulation study and the analysis of real-life call center data. Mandelbaum and Zeltyn (2010) advocated a process-view of service times which are modeled as the evolution of a finite-state continuous-time absorbing Markov process (phasetype distribution). Here, even though we use additional information when modelling service times, such as the time when the call is answered, we continue to assume the lognormality of the individual service times.…”

Active Learning in Fashion and Textiles Supply Chain Management

“…As a result, another future research area is to investigate how analytical models can be incorporated into the teaching and learning materials for fashion and textiles supply chain management under the scenario that the students have very diversified backgrounds. The reference work by Mandelbaum and Zeltyn [9] is very useful for this new research direction. 4.…”