EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
2666. Learning and pricing in queues with unobserved balking
Invited abstract in session TA-40: Queueing Models with Strategic Customers, stream Advances in Stochastic Modelling and Learning Methods.
Tuesday, 8:30-10:00Room: 96 (building: 306)
Authors (first author is the speaker)
| 1. | Shreehari Bodas
|
| Korteweg-de Vries Institute for Mathematics, University of Amsterdam | |
| 2. | Michel Mandjes
|
| University of Amsterdam | |
| 3. | Liron Ravner
|
| Statistics, University of Haifa |
Abstract
Impatience of customers plays a crucial role in modelling service systems. In many situations, any arriving customer is informed about the current congestion level. In this way they can obtain a sense of their prospective waiting time, based on which they can decide whether or not to join. If this waiting time exceeds their patience threshold, they balk; otherwise they join. In this talk we discuss queues with impatient customers, by in particular considering a relevant related decision problem: how to set a fixed admission price so as to maximize revenues? We discuss approaches based on applying existing techniques such as Stochastic Gradient Descent via Infinitesimal Perturbation Analysis and Stochastic Approximation methods such as the Kiefer-Wolfowitz algorithm.
Keywords
- Queuing Systems
- Stochastic Models
- Simulation
Status: accepted
Back to the list of papers