TY - GEN
T1 - External-cost continuous-type wardrop equilibria in routing games
AU - Calderone, Dan
AU - Dong, Roy
AU - Sastry, S. Shankar
N1 - Funding Information:
This work is supported by NSF FORCES CNS-1239166, AFOSR MURI CHASE FA9550-1 0-1-0567, and ONR N00014-09-1-0230.
Publisher Copyright:
© 2017 IEEE.
PY - 2018/3/14
Y1 - 2018/3/14
N2 - We propose a bi-criterion Wardrop equilibrium which we call an external cost continuous type Wardrop equilibrium for nonatomic routing games where agents incur some type dependent cost to have access to different sets of available routes. Rather than being anonymous, the population is described by a distribution over the type parameter (that can be supported on any compact interval of the real line). At equilibria, no member of the population can improve their type dependent cost either by switching sets of routes or by switching routes within their chosen set. From this equilibrium condition, we derive how the population mass will divide up among the various routing options. We then formulate a potential function optimization program for finding the equilibrium mass distribution. This work revisits the cost-vs-time equilibria of Leurent [1] and Marcotte [2] while specifically allowing the type parameter to be positive or negative. Applications include modeling the value of information in routing, the effect of privacy concern on congestion, and how commuters make tradeoffs between different forms of transportation.
AB - We propose a bi-criterion Wardrop equilibrium which we call an external cost continuous type Wardrop equilibrium for nonatomic routing games where agents incur some type dependent cost to have access to different sets of available routes. Rather than being anonymous, the population is described by a distribution over the type parameter (that can be supported on any compact interval of the real line). At equilibria, no member of the population can improve their type dependent cost either by switching sets of routes or by switching routes within their chosen set. From this equilibrium condition, we derive how the population mass will divide up among the various routing options. We then formulate a potential function optimization program for finding the equilibrium mass distribution. This work revisits the cost-vs-time equilibria of Leurent [1] and Marcotte [2] while specifically allowing the type parameter to be positive or negative. Applications include modeling the value of information in routing, the effect of privacy concern on congestion, and how commuters make tradeoffs between different forms of transportation.
UR - http://www.scopus.com/inward/record.url?scp=85046269113&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85046269113&partnerID=8YFLogxK
U2 - 10.1109/ITSC.2017.8317866
DO - 10.1109/ITSC.2017.8317866
M3 - Conference contribution
AN - SCOPUS:85046269113
T3 - IEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC
SP - 1
EP - 6
BT - 2017 IEEE 20th International Conference on Intelligent Transportation Systems, ITSC 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 20th IEEE International Conference on Intelligent Transportation Systems, ITSC 2017
Y2 - 16 October 2017 through 19 October 2017
ER -