Service Operations

Many of the research activities of the stochastic modeling group fall within the broad area of service systems. Some of these activities are described as part of other more focused research projects (e.g., patient flow, appointment scheduling and healthcare capacity, queueing design and control etc.) but the group is also involved in research that either aims to make a broader contribution to the analysis of service systems or to model and analyze specialized service systems that have not attracted much academic attention before. Below are a few examples:

Kulkarni is interested in the cash replenishment decisions for Automatic Teller Machines (ATMs), which are used to provide basic banking services, such as cash withdrawals and deposits. The picture below shows the ATMs owned by SECU in the triangle area. (ATM symbol stands for stand-alone ATM, and a $ symbol stands for an ATM in a bank). The ATM fees is one of the largest sources of income for SECU. One needs to keep the ATMS properly filled with cash so that they can operate 24-7. The replenishment operation involves replacing the current cash module in the ATM with a new module full of cash. The replenishment policy decides what set of ATMS should be replaced, based on the current cash level at all the ATMs. The stochastic model involves statistical analysis of the cash demands at ATMs to create an accurate forecast, traveling salesman type analysis to determine the routing costs, and optimization of submodular function to identify the optimal set of ATMs to replenish.

 

Ziya is interested in optimal control and design of service systems when customers make their own strategic decisions. For example, in many service systems where customers wait to receive service, service speed is not known with certainty. Arriving customers would choose to join or not depending on their rough estimate for their waiting time. As they wait, they would have a better idea about the service speed and periodically review their decision to continue to wait or simply leave the queue. In some other settings where service is a shared experience (e.g. clubs, gyms etc.) customers might have some sensitivity about the number and type of other customers. The control and design decisions for such systems have not been adequately studied and some of Ziya’s work contributes to filling that gap.

Students who have been involved in these research projects include Yu Zhang and Yichen Tu (Ph.D. students). If you are interested in being involved in this research project, please send an email to Argon, Kulkarni or Ziya for any potential opportunities.

Selected Publications:

1 – Zhang, Y., V. G. Kulkarni, “Automated Teller Machines Replenishment Policies with Submodular Costs,” forthcoming in Manufacturing and Service Operations Management.

2 – Kostami, V., Kostamis, D., and S. Ziya, “Pricing and capacity allocation for shared services,” Manufacturing and Service Operations Management, 19 (2017) 230-245. (supplement)