Deterministic state-based information disclosure policies and social welfare maximization in strategic queueing systems
- PDF / 688,366 Bytes
- 26 Pages / 439.37 x 666.142 pts Page_size
- 35 Downloads / 168 Views
Deterministic state-based information disclosure policies and social welfare maximization in strategic queueing systems Tesnim Naceur1
· Yezekael Hayel1
Received: 28 October 2019 / Revised: 1 October 2020 / Accepted: 13 October 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Understanding the impact of the queue length information on the behavior of customers is nowadays a hot topic for strategic queueing problems. Particularly, it is important to understand how this information impacts the equilibrium behavior and therefore global performance metrics in order to give some recommendations to service providers for optimizing their objective. Recent works are focused on information policies in order to optimize the provider’s profit. In this work, the aim is to maximize the expected social welfare. It is intuitive that the expected social welfare is higher when more information is available to new customers. In some circumstances, and particularly when the queue is overloaded, providing the information up to a particular level of occupancy yields a better expected social welfare compared to always providing queue length information to new customers. This counterintuitive result comes mainly from the strategic behavior of customers at equilibrium when queue length information is not available. After proving that there exists such an optimal information disclosure policy (IDP) for arbitrary arrival and service rates, we go further and study the expected social welfare for the overloaded context in which the arrival rate equals the service rate. Finally, numerical illustrations corroborate our contributions, and moreover, they show that an admission control policy can be also proposed, combined with an IDP, in order to optimize the expected social welfare of the system. Keywords Information in strategic queueing · Social welfare optimization · Information disclosure policies
The authors acknowledge support from CEFIPRA 5702-1 Project for this work.
B
Tesnim Naceur [email protected] Yezekael Hayel [email protected]
1
LIA/CERI, University of Avignon, Avignon, France
123
Queueing Systems
Mathematics Subject Classification 68M20 · 91A40
1 Introduction Strategic decisions in queueing systems, also called strategic queueing, gain lot of interest in the design of complex systems in which many decision makers/customers interact. In fact, taking into account how customers react to signals like access price or queue length information is crucial in order to optimize the global behavior of such systems in which the interaction structure can be modeled as a queueing system. Large complex systems involve so many customers such that an infinite or a population of strategic individuals can be considered. Moreover, customers interact through a queuing system and the waiting/sojourn time is the main performance metric incurred by them. 1.1 Information in strategic queueing systems Applications of such strategic complex systems are broadly in the field of “Ne
Data Loading...
