Traditional queueing models for service systems assume that customer characteristics (such as service requirement and patience) are “assigned” to each customer independently of all other random variables describing the system, so that customers are mathematically treated as inanimate jobs. However, an important aspect in reality, that is mostly ignored, is that the service requirement of a customer depends on her patience, or on the delay she experienced in queue. Such dependencies clearly hold in healthcare settings (e.g., for stroke patients), and have also been empirically shown to exist in other important systems, such as contact centers and restaurants. The goal of the proposed work is to design models for systems with such dependencies, and to develop the required tools to analyze them. Specifically, two different types of dependencies are considered: Dependence between the service and patience times of a customer, and dependence of the service requirement on the delay in queue. The former dependence is referred to as “exogenous,” because it is an attribute of the customer (and thus ``arrives’’ to the system from the outside), and the latter as “endogenous,” because the dependence is endogenized by the system’s dynamics. It is shown that, when either type of dependence exists, it must be considered, because ignoring it leads to significant quantitative and qualitative errors. Since exact analysis is intractable even for the simplest systems, various approximations are proposed, and preliminary investigations demonstrate that those approximations are accurate and effective.
|Effective start/end date||9/1/20 → 8/31/23|
- National Science Foundation (CMMI-2006350)
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.