Reducing the maximum latency of selfish ring routing via pairwise cooperations

Authors:
Xujin Chen;Xiaodong Hu;Weidong Ma
Affiliations:
Institute of Applied Mathematics, Chinese Academy of Sciences, Beijing, China;Institute of Applied Mathematics, Chinese Academy of Sciences, Beijing, China;Institute of Applied Mathematics, Chinese Academy of Sciences, Beijing, China
Venue:
COCOA'10 Proceedings of the 4th international conference on Combinatorial optimization and applications - Volume Part II
Year:
2010

Citing 9
Cited 1

The Ring Loading Problem

SIAM Journal on Discrete Mathematics
How bad is selfish routing?

Journal of the ACM (JACM)
The price of anarchy is independent of the network topology

Journal of Computer and System Sciences - STOC 2002
Linear time algorithms for the ring loading problem with demand splitting

Journal of Algorithms
The Price of Routing Unsplittable Flow

Proceedings of the thirty-seventh annual ACM symposium on Theory of computing
The price of anarchy of finite congestion games

Proceedings of the thirty-seventh annual ACM symposium on Theory of computing
Algorithmic Game Theory

Algorithmic Game Theory
Path decomposition under a new cost measure with applications to optical network design

ACM Transactions on Algorithms (TALG)
Worst-case equilibria

STACS'99 Proceedings of the 16th annual conference on Theoretical aspects of computer science

Selfish routing and wavelength assignment strategies with advance reservation in inter-domain optical networks

Computer Communications

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper studies the selfish routing game in ring networks with a load-dependent linear latency on each link. We adopt the asymmetric atomic routing model. Each player selfishly chooses a route to connect his source-destination pair, aiming at a lowest latency of his route, while the system objective is to minimize the maximum latency among all routes of players. Such a routing game always has a Nash equilibrium (NE) that is a "stable state" among all players, from which no player has the incentive to deviate unilaterally. Furthermore, 16 is the current best upper bound on its price of anarchy (PoA), the worst-case ratio between the maximum latencies in a NE and in a system optimum. In this paper we show that the PoA is at most 10.16 provided cooperations within pairs of players are allowed, where any two players could change their routes simultaneously if neither would experience a longer latency and at least one would experience a shorter latency.