Doctoral thesis

Study on behavioral patterns in queuing : Agent based modeling and experimental approach

    27.10.2011

117 p

Thèse de doctorat: Università della Svizzera italiana, 2011 (jury note: summa cum laude)

English “Time is money” if you are a service provider or a customer using the service. In the current economic scenario where there is heavy competition, customers have more alternatives at their disposal; which means understanding customers is of paramount importance. This dissertation delves into this terrain, and tries to provide a conceptual and an experimental framework to understand the complex dynamics we see every day in queuing. The traditional approach to queuing has been towards designing an efficient facility, with little focus towards the customers who use these facilities. This approach has been helpful to understand the capacity requirements, and the impact of different configurations of production and service facilities, but has limited our ability to explain the behavior observed in many real queues. This dissertation continues the shift in focus from the service facility design to the information structure, and the behavior of the customers in a queuing system. Two extensions are proposed to the traditional queuing framework: Customers are provided with 1) information regarding the performance of the facility (feedback) and 2) computational capability to process the information provided. In a nutshell, the customers in this model choose which facility to use depending on their expectations which are based on past experiences. I present two agent based modeling frameworks to characterize the customers using the facility. This approach helps us to model homogenous customers, and it is useful because every customer perceives differently the cost of waiting and also the efficiency of the system. Agent based models help to understand how the customers might co-evolve with the facility that serves them, which is difficult to achieve within the traditional queuing framework. Finally, an experiment is designed using the guidelines from experimental economics to validate the agent based modeling framework.
Language
  • English
Classification
Economics
License
License undefined
Identifiers
Persistent URL
https://n2t.net/ark:/12658/srd1318243
Statistics

Document views: 74 File downloads:
  • 2011ECO006.pdf: 118