628. Optimal and approximate solutions to the backup server selection problem
Invited abstract in session WA-54: Applications in Queueing Theory, stream Stochastic modelling.
Wednesday, 8:30-10:00Room: Liberty 1.08
Authors (first author is the speaker)
| 1. | Vasco Alves
|
| Department of Economics, University of Birmingham | |
| 2. | Refael Hassin
|
| Statistics, Tel Aviv University | |
| 3. | Jiesen Wang
|
| Tel Aviv University |
Abstract
Backup servers are often used in service systems when the primary servers are overloaded. Backup servers may have a lower service quality or a slower service rate. We analyze a system with one primary and one backup server and compute the system and individually optimal server-selection strategies. We also compare the optimal strategy with the better of two simpler strategies: using only the primary server or never waiting for the primary server when the backup server is idle. We demonstrate numerically that the better of two simpler strategies is a good approximation, in terms of optimal social welfare. We also compare the decentralized equilibrium strategy to the socially-optimal behavior. We show that the Price of Anarchy can be arbitrarily large. However, the socially-optimal behavior can be induced by charging an admission fee to the system and a price for using the high quality server.
Keywords
- Queuing Systems
- Game Theory
- Stochastic Models
Status: accepted
Back to the list of papers