#### DMCA

## How bad is selfish routing? (2002)

### Cached

### Download Links

- [www.cs.ust.hk]
- [www.cs.cuhk.hk]
- [www.cse.cuhk.edu.hk]
- [www.cs.cuhk.edu.hk]
- [www.cs.cmu.edu]
- [www-2.cs.cmu.edu]
- [www.cs.cornell.edu]
- [www-2.cs.cmu.edu]
- [www.cs.cmu.edu]
- [www.cs.cmu.edu]
- [www.cs.cmu.edu]
- [www.cs.huji.ac.il]
- [www.cs.northwestern.edu]
- [personal.ie.cuhk.edu.hk]
- [www.cs.cornell.edu]
- [lcm.csa.iisc.ernet.in]
- [www.cs.cornell.edu]
- [www.cs.stanford.edu]
- [theory.stanford.edu]
- [theory.stanford.edu]
- [www.cs.cornell.edu]
- [www.cs.brown.edu]
- [theory.stanford.edu]
- [www.cs.stanford.edu]
- [www.cs.cornell.edu]
- [cs.uwaterloo.ca]
- [theory.stanford.edu]
- [cs.uwaterloo.ca]
- [theory.stanford.edu]
- [theory.stanford.edu]

Venue: | JOURNAL OF THE ACM |

Citations: | 652 - 27 self |

### Citations

2869 | The Design and Analysis of Computer I Algorithms - Aho, Hopcroft, et al. - 1974 |

2335 | Random Graphs - Bollobás - 2001 |

2192 | Data Networks
- Bertsekas, Gallager
- 1992
(Show Context)
Citation Context ...n arbitrarily large amount. This example is of particular interest because functions of the form ℓ(x) =1/(u−x) arise as the delay functions of M/M/1 queues with capacity or service rate u (see, e.=-=g., [4]-=-) and are therefore common in the networking literature [4, 26, 27, 29, 35]. However, in networks where the largest possible change in edge latency resulting from a single agent rerouting its flow is ... |

1981 | Network Flows: Theory, Algorithms and Application - Ahuja, Magnanti, et al. - 1993 |

830 | Worst-case equilibria
- Koutsoupias, Papadimitriou
(Show Context)
Citation Context ...Our approach to quantifying the inefficiency inherent in a selfishly defined solution (dubbed the “price of anarchy” by Papadimitriou [37]) is inspired by recent work of Koutsoupias and Papadimitriou =-=[28]-=-. In [28], network routing was modeled as a noncooperative game (though with a different model than ours, and only for two-node networks) and the worst-case ratio of the social welfare achieved by a N... |

755 | The Linear Complementarity Problem - Cottle, Pang, et al. - 1993 |

751 |
Some theoretical aspects of road traffic research
- Wardrop
- 1952
(Show Context)
Citation Context ...f our results to these assumptions. Related Work Unregulated traffic has been modeled as network flow with all flow paths between a given source-destination pair having equal latency since the 1950’s =-=[3, 52]-=- (see also Knight [24]). Beckman et al. [3], observing that such an equilibrium flow is an optimal solution to a related convex program (see also Section 2), gave existence and uniqueness results for ... |

741 | Dynamic Noncooperative Game Theory - Basar, Olsder - 1998 |

556 |
A class of games possessing pure-strategy nash equilibria
- Rosenthal
- 1973
(Show Context)
Citation Context ...Braess’s Paradox. The addition of an intuitively helpful link can negatively impact all of the users of a congested network. of and extensions to this traffic model have been studied (see for example =-=[1, 11, 16, 21, 32, 33, 34, 35, 41, 43, 47, 49]-=-). In the past several decades, much of the work on this traffic model has been inspired by a “paradox” first discovered by Braess [6] and later reported by Murchland [31] (see also [2] for a non-tech... |

415 |
Existence and uniqueness of equilibrium points for concave N-Person games
- Rosen
- 1965
(Show Context)
Citation Context ...ibrium if and only if for each i, f (i) minimizes Ci(f) givenf (j) for j �= i. We will focus on the case where for each edge e, x · ℓe(x) is a convex function; under this assumption, results of Rosen =-=[40]-=- imply that a flow at Nash equilibrium must exist and will be essentially unique. Our main result for this model is an analogue of Theorem 3.1. 19sTheorem 5.4 If f is at Nash equilibrium for the finit... |

377 |
Game Theory
- Owen
- 1995
(Show Context)
Citation Context ...umption, we can view network users as independent agents participating in a non-cooperative game and expect the routes chosen by users to form a Nash equilibrium in the sense of classical game theory =-=[24]-=-. In other words, we assume that each agent uses the minimumlatency path from its source to its destination, given the link congestion caused by the rest of the network users. It is well-known that Na... |

334 |
the Internet
- Algorithms
- 2001
(Show Context)
Citation Context ...ilibrium with that of the optimal assignment of traffic to paths. Our approach to quantifying the inefficiency inherent in a selfishly defined solution (dubbed the “price of anarchy” by Papadimitriou =-=[37]-=-) is inspired by recent work of Koutsoupias and Papadimitriou [28]. In [28], network routing was modeled as a noncooperative game (though with a different model than ours, and only for two-node networ... |

229 | Building Secure and Reliable Network Applications - Birman - 1996 |

221 | Competitive routing in multi‐user communication networks
- Orda, Rom, et al.
- 1993
(Show Context)
Citation Context ...Braess’s Paradox. The addition of an intuitively helpful link can negatively impact all of the users of a congested network. of and extensions to this traffic model have been studied (see for example =-=[1, 11, 16, 21, 32, 33, 34, 35, 41, 43, 47, 49]-=-). In the past several decades, much of the work on this traffic model has been inspired by a “paradox” first discovered by Braess [6] and later reported by Murchland [31] (see also [2] for a non-tech... |

217 | Urban Transportation Networks: Equilibrium Analysis with Mathematical Programming Methods - Sheffi - 1985 |

212 | The Price of Anarchy is Independent of the Network Topology
- Roughgarden
- 2002
(Show Context)
Citation Context ...e prove that if (G, r, ℓ) is an instance with linear latency functions, then ρ(G, r, ℓ) ≤ 4. The upper bound of Corollary 2.8 is not best possible for higher-degree 3 polynomials, either; Roughgarden =-=[44]-=- has recently shown that if all latency functions of an instance (G, r, ℓ) are polynomials with degree at most p, thenρ(G, r, ℓ) =O( p ln p )(see Section 6 for details). On the other hand, for each va... |

207 | Speed is as powerful as clairvoyance
- Kalyanasundaram, Pruhs
(Show Context)
Citation Context ...16, 17] for a survey). Since these early works, many properties 2 This approach is thus in the spirit of the analyses of online algorithms via resource augmentation given by Kalyanasundaram and Pruhs =-=[23] a-=-nd Phillips et al. [39]. 3sl (x) = x v s t l (x) = 1 w (a) Before l (x) = 1 l (x) = x l (x) = x v s l (x) = 0 t l(x) = 1 l(x) = x w (b) After l (x) = 1 Figure 1: Braess’s Paradox. The addition of an... |

198 | Network Optimization: Continuous and Discrete Models, Athena Scientific - Bertsekas - 1998 |

177 |
The traffic assignment problem for a general network
- Dafermos, Sparrow
- 1969
(Show Context)
Citation Context ...3], observing that such an equilibrium flow is an optimal solution to a related convex program (see also Section 2), gave existence and uniqueness results for traffic equilibria. Dafermos and Sparrow =-=[13]-=- were perhaps the first authors interested in computing the equilibrium efficiently, and many subsequent papers gave increasingly efficient methods for computing equilibria (see [16, 17] for a survey)... |

173 | Tight Bounds for Worst-case Equilibria
- Czumaj, Vöcking
- 2002
(Show Context)
Citation Context ...three or more parallel links; tight results have recently been obtained for parallel networks with any number of links by Mavronicolas and Spirakis [30] (for a special case) and by Czumaj and Vöcking =-=[10]-=- (for the general case). Equilibria in Other Settings Braess’s Paradox is not particular to traffic in networks; perhaps the most compelling analogue occurs in a mechanical network of strings and spri... |

153 | Optimal time-critical scheduling via resource augmentation
- Phillips, Stein, et al.
(Show Context)
Citation Context ...nce these early works, many properties 2 This approach is thus in the spirit of the analyses of online algorithms via resource augmentation given by Kalyanasundaram and Pruhs [23] and Phillips et al. =-=[39]-=-. 3sl (x) = x v s t l (x) = 1 w (a) Before l (x) = 1 l (x) = x l (x) = x v s l (x) = 0 t l(x) = 1 l(x) = x w (b) After l (x) = 1 Figure 1: Braess’s Paradox. The addition of an intuitively helpful link... |

153 | Recommendations on queue management and congestion avoidance in the internet
- Braden, Clark, et al.
- 1998
(Show Context)
Citation Context ...network performance. For example, existing Internet protocols place little restriction on how network traffic is routed, allowing network users to make decisions in a selfish or even malicious manner =-=[3]-=-. The central question of this paper is how much does network performance suffer from this lack of regulation ? As a first step toward formalizing this question mathematically, we assume that, in the ... |

132 | The price of selfish routing
- Mavronicolas, Spirakis
- 2001
(Show Context)
Citation Context ...works and partial results for twonode networks with three or more parallel links; tight results have recently been obtained for parallel networks with any number of links by Mavronicolas and Spirakis =-=[30]-=- (for a special case) and by Czumaj and Vöcking [10] (for the general case). Equilibria in Other Settings Braess’s Paradox is not particular to traffic in networks; perhaps the most compelling analogu... |

131 |
Existence, uniqueness and stability of traffic equilibria
- Smith
- 1979
(Show Context)
Citation Context ...Braess’s Paradox. The addition of an intuitively helpful link can negatively impact all of the users of a congested network. of and extensions to this traffic model have been studied (see for example =-=[1, 11, 16, 21, 32, 33, 34, 35, 41, 43, 47, 49]-=-). In the past several decades, much of the work on this traffic model has been inspired by a “paradox” first discovered by Braess [6] and later reported by Murchland [31] (see also [2] for a non-tech... |

120 | Stackelberg Scheduling Strategies
- Roughgarden
- 2001
(Show Context)
Citation Context |

111 | A survey on networking games in telecommunications - Altman, Boulogne, et al. |

106 |
Some fallacies in the interpretation of social cost
- Knight
- 1924
(Show Context)
Citation Context ...sumptions. Related Work Unregulated traffic has been modeled as network flow with all flow paths between a given source-destination pair having equal latency since the 1950’s [3, 52] (see also Knight =-=[24]-=-). Beckman et al. [3], observing that such an equilibrium flow is an optimal solution to a related convex program (see also Section 2), gave existence and uniqueness results for traffic equilibria. Da... |

93 |
Traffic equilibrium and variational inequalities
- Dafermos
- 1980
(Show Context)
Citation Context |

93 |
On the relationship between nash cournot and wardrop equilibria
- Haurie, Marcotte
- 1985
(Show Context)
Citation Context .... fP Letting δ tend to 0, continuity and monotonicity of the edge latency functions give the following useful characterization of a flow at Nash equilibrium, occasionally called a Wardrop equilibrium =-=[22]-=- or Wardrop’s Principle [50, 51] in the literature, due to an influential paper of Wardrop [52]. Lemma 2.2 Aflowf feasible for instance (G, r, ℓ) is at Nash equilibrium if and only if for every i ∈{1,... |

88 | Game Theory, 3rd Ed - Owen - 2001 |

87 | Competitive routing in networks with polynomial costs - Altman, Basar, et al. - 2002 |

85 |
Inefficiency of Nash equilibria
- Dubey
- 1986
(Show Context)
Citation Context ...ongestion caused by the rest of the network users. It is well-known that Nash equilibria do not in general optimize social welfare; perhaps the most famous example is that of “The Prisoner’s Dilemma” =-=[12, 28]-=-. We are then interested in comparing the total latency of a Nash equilibrium with that of the optimal assignment of traffic to paths. This line of research was recently initiated by Koutsoupias and P... |

77 | Capacity allocation under noncooperative routing
- Korilis, Lazar, et al.
- 1997
(Show Context)
Citation Context ...Paradox (even in its worst-possible manifestation) is algorithmically difficult [42]. In a related model with finitely many agents, each controlling a strictly positive amount of flow, Korilis et al. =-=[26, 27]-=- studied strategies for adding new edges and/or capacity to a network that guarantee an improvement in network performance. In contrast to this previous work, we are interested in quantifying the diff... |

70 | Equilibria on a Congested Transportation Network
- Aashtiani, Magnanti
- 1981
(Show Context)
Citation Context |

65 | Designing networks for selfish users is hard
- Roughgarden
- 2001
(Show Context)
Citation Context ...work performance [18, 51], discovered new types of “paradoxes” [12, 15, 20, 48, 50], and proved that detecting Braess’s Paradox (even in its worst-possible manifestation) is algorithmically difficult =-=[42]-=-. In a related model with finitely many agents, each controlling a strictly positive amount of flow, Korilis et al. [26, 27] studied strategies for adding new edges and/or capacity to a network that g... |

62 | Avoiding the Braess paradox in noncooperative networks
- Korilis, Lazar, et al.
- 1999
(Show Context)
Citation Context ...Paradox (even in its worst-possible manifestation) is algorithmically difficult [42]. In a related model with finitely many agents, each controlling a strictly positive amount of flow, Korilis et al. =-=[26, 27]-=- studied strategies for adding new edges and/or capacity to a network that guarantee an improvement in network performance. In contrast to this previous work, we are interested in quantifying the diff... |

60 |
transportation networks: Equilibrium analysis with mathematical programming methods
- Urban
- 1985
(Show Context)
Citation Context ...ntly, and many subsequent papers gave increasingly efficient methods for computing equilibria (see [10] for a survey); others have extended these results to more sophisticated models (see for example =-=[1, 6, 10, 13, 21, 22, 27, 29, 30]). In the -=-past several decades, much of the work on this model has been inspired by a "paradox" first discovered by Braess [4] and later reported by Murchland [20]. The essence of Braess's Paradox is ... |

57 |
A paradox of congestion in a queuing network
- Cohen, Kelly
- 1990
(Show Context)
Citation Context ...hibited in two-terminal electrical networks [7] (where our results give analogous bounds on the largest possible increase in conductivity obtainable by removing conducting links) and queuing networks =-=[8]-=-. Organization In Section 2 we give a formal definition of our network model and of flows at Nash equilibrium, and state several lemmas needed for our main results. In Section 3 we prove our main bicr... |

56 |
Uber ein Paradoxon der Verkehrsplanung,” Unternehmenforschung 12
- Braess
- 1968
(Show Context)
Citation Context ...died (see for example [1, 11, 16, 21, 32, 33, 34, 35, 41, 43, 47, 49]). In the past several decades, much of the work on this traffic model has been inspired by a “paradox” first discovered by Braess =-=[6]-=- and later reported by Murchland [31] (see also [2] for a non-technical account). The essence of Braess’s Paradox is captured by the example shown in Figure 1, where the edges are labeled with their l... |

56 |
Studies in the Economics of Transportation
- Beckmann, McGuire, et al.
- 1956
(Show Context)
Citation Context ...tination have equal (and smallest possible) latency—if a flow does not have this property, some agent can improve its travel time by switching from a longer flow path to a shorter one. Beckman et al=-=. [3]-=- showed that if the latency of each network link is a continuous nondecreasing function of the flow on the link, then a flow corresponding to a Nash equilibrium always exists and moreover all such flo... |

51 |
Network equilibrium models and algorithms
- Florian, Hearn
- 1995
(Show Context)
Citation Context ...ermos and Sparrow [13] were perhaps the first authors interested in computing the equilibrium efficiently, and many subsequent papers gave increasingly efficient methods for computing equilibria (see =-=[16, 17]-=- for a survey). Since these early works, many properties 2 This approach is thus in the spirit of the analyses of online algorithms via resource augmentation given by Kalyanasundaram and Pruhs [23] an... |

48 |
Braess’ Paradox of Traffic Flow
- Murchland
- 1970
(Show Context)
Citation Context ..., 32, 33, 34, 35, 41, 43, 47, 49]). In the past several decades, much of the work on this traffic model has been inspired by a “paradox” first discovered by Braess [6] and later reported by Murchland =-=[31]-=- (see also [2] for a non-technical account). The essence of Braess’s Paradox is captured by the example shown in Figure 1, where the edges are labeled with their latency functions (each a function of ... |

46 |
The mathematics of nonlinear programming
- Peressini, Sullivan, et al.
- 1988
(Show Context)
Citation Context ...s optimal for a convex program of the form (NLP) if and only if for every i ∈{1,...,k} and P1,P2 ∈Pi with fP1 > 0, c ′ P1 (f) ≤ c′ P2 (f). 4 For a formal derivation via the Karush-Kuhn-Tucker Theorem =-=[38]-=-, see [3, 13]. 8sRemark: To see that this characterization still makes sense under the weaker assumption of continuous (and not necessarily differentiable) latency functions, define c− e (x),c+ e (x) ... |

41 |
On Some Traffic Equilibrium Theory Paradoxes
- Dafermos, Nagurney
- 1984
(Show Context)
Citation Context ...w roads in congested cities [25, 31], researchers attempted to classify networks in which the addition of a single link could degrade network performance [18, 51], discovered new types of “paradoxes” =-=[12, 15, 20, 48, 50]-=-, and proved that detecting Braess’s Paradox (even in its worst-possible manifestation) is algorithmically difficult [42]. In a related model with finitely many agents, each controlling a strictly pos... |

41 | Virtual path bandwidth allocation in multiuser networks
- Lazar, Orda, et al.
- 1997
(Show Context)
Citation Context ...nterest because functions of the form ℓ(x) =1/(u−x) arise as the delay functions of M/M/1 queues with capacity or service rate u (see, e.g., [4]) and are therefore common in the networking literature =-=[4, 26, 27, 29, 35]-=-. However, in networks where the largest possible change in edge latency resulting from a single agent rerouting its flow is not too large, we can apply the results of Subsection 5.1 to derive the fol... |

41 | Congestion resulting from increased capacity in sigleserver queueing networks - Cohen, Jeffries - 1997 |

39 | Optimal Time-Varying Flows on Congested Networks - Carey - 1987 |

31 | Routing in two parallel links: Game-theoretic distributed algorithms. to appear in - Altman, Basar, et al. - 2001 |

30 |
The Braess paradox
- Frank
- 1981
(Show Context)
Citation Context ...a flow minimizing the total latency)? Our Results In networks in which the latency of each edge is a linear function of the edge congestion (a model that has been the focus of several previous papers =-=[7, 18, 51]-=-), we show that a flow at Nash equilibrium has total latency at most 4 times that of the optimal flow. We give 3 examples showing that this result is tight. We also consider the model in which link la... |

28 | Braess’s paradox in a queueing network with state-dependent routing - Calvert, Solomon, et al. - 1997 |

25 |
Nonlinear Cost Network Models in Transportation Analysis
- Florian
- 1986
(Show Context)
Citation Context ...ermos and Sparrow [13] were perhaps the first authors interested in computing the equilibrium efficiently, and many subsequent papers gave increasingly efficient methods for computing equilibria (see =-=[16, 17]-=- for a survey). Since these early works, many properties 2 This approach is thus in the spirit of the analyses of online algorithms via resource augmentation given by Kalyanasundaram and Pruhs [23] an... |

25 | Equilibria for multiclass routing in multi-agent networks - Altman, Kameda - 2001 |

25 | Mixed equilibrium for multiclass routing games - Boulogne, Altman, et al. - 2002 |

24 |
More Paradoxes in the Equilibrium Assignment Problem
- Fisk
- 1979
(Show Context)
Citation Context ...w roads in congested cities [25, 31], researchers attempted to classify networks in which the addition of a single link could degrade network performance [18, 51], discovered new types of “paradoxes” =-=[12, 15, 20, 48, 50]-=-, and proved that detecting Braess’s Paradox (even in its worst-possible manifestation) is algorithmically difficult [42]. In a related model with finitely many agents, each controlling a strictly pos... |

24 |
Some theoretical aspects of road trac research
- Wardrop
- 1952
(Show Context)
Citation Context ... of our results to these assumptions. Related Work Unregulated tra#c has been modeled as network flow with all flow paths between a given source-destination pair having equal latency since the 1950's =-=[3, 52]-=- (see also Knight [24]). Beckman et al. [3], observing that such an equilibrium flow is an optimal solution to a related convex program (see also Section 2), gave existence and uniqueness results for ... |

23 | Traffic equilibrium paradoxes - Catoni, Pallottino - 1991 |

20 |
Graphentheoretische Methoden und ihre Anwendungen
- Knodel
- 1969
(Show Context)
Citation Context ...he agents! Motivated by the discovery of Braess’s Paradox and evidence of similarly counterintuitive and counterproductive traffic behavior following the construction of new roads in congested cities =-=[25, 31]-=-, researchers attempted to classify networks in which the addition of a single link could degrade network performance [18, 51], discovered new types of “paradoxes” [12, 15, 20, 48, 50], and proved tha... |

20 |
The prevalence of Braess’ paradox
- Steinberg, Zangwill
- 1983
(Show Context)
Citation Context ...a flow minimizing the total latency)? Our Results In networks in which the latency of each edge is a linear function of the edge congestion (a model that has been the focus of several previous papers =-=[7, 18, 51]-=-), we show that a flow at Nash equilibrium has total latency at most 4 times that of the optimal flow. We give 3 examples showing that this result is tight. We also consider the model in which link la... |

19 |
Properties of the equilibrium state in transportation networks
- Hall
- 1978
(Show Context)
Citation Context |

19 |
Sustainable Transportation Networks
- Nagurney
- 2000
(Show Context)
Citation Context |

19 | Braess’s paradox in a loss network - Bean, Kelly, et al. - 1997 |

19 | Anonymous Games with Binary Actions - Blonski - 1999 |

19 | Stackelberg Strategies for Multilevel Systems - Cruz - 1976 |

13 |
On the algebra of network
- Bott, Duffin
(Show Context)
Citation Context ...roportional to its congestion; this result is implicit in the work of Dafermos and Sparrow [13], and other properties of this special case have been investigated in the context of electrical networks =-=[5, 9]-=-. Corollary 4.2 Let G be a network in which each edge latency function ℓe is of the form ℓe(x) =aex. Then for any rate vector r, aflowfeasiblefor(G, r, ℓ) is optimal if and only if it is at Nash equil... |

12 | Braess’s paradox and power-law nonlinearities in networks, Part II - Calvert, Keady - 1995 |

11 |
In a road network, increasing delay locally can reduce delay globally
- Smith
- 1978
(Show Context)
Citation Context ...w roads in congested cities [25, 31], researchers attempted to classify networks in which the addition of a single link could degrade network performance [18, 51], discovered new types of “paradoxes” =-=[12, 15, 20, 48, 50]-=-, and proved that detecting Braess’s Paradox (even in its worst-possible manifestation) is algorithmically difficult [42]. In a related model with finitely many agents, each controlling a strictly pos... |

9 | Characterizing Braess's Paradox for Traffic Networks
- Hagstrom
- 2001
(Show Context)
Citation Context |

8 | Braess’s Paradox of Traffic Flow,” Transportation Research - Murchland - 1970 |

7 |
Road to ruin
- Bass
- 1992
(Show Context)
Citation Context ...5, 41, 43, 47, 49]). In the past several decades, much of the work on this traffic model has been inspired by a “paradox” first discovered by Braess [6] and later reported by Murchland [31] (see also =-=[2]-=- for a non-technical account). The essence of Braess’s Paradox is captured by the example shown in Figure 1, where the edges are labeled with their latency functions (each a function of the link conge... |

7 | Road to Ruin,” Discover - Bass - 1992 |

6 |
Paradoxical behavior of mechanical and electrical networks. Nature 352
- Cohen, Horwitz
- 1991
(Show Context)
Citation Context ...a flow minimizing the total latency)? Our Results In networks in which the latency of each edge is a linear function of the edge congestion (a model that has been the focus of several previous papers =-=[7, 18, 51]-=-), we show that a flow at Nash equilibrium has total latency at most 4 times that of the optimal flow. We give 3 examples showing that this result is tight. We also consider the model in which link la... |

6 |
On some trac equilibrium theory paradoxes," Transportation Research 18B
- Dafermos, Nagurney
- 1984
(Show Context)
Citation Context ...w roads in congested cities [25, 31], researchers attempted to classify networks in which the addition of a single link could degrade network performance [18, 51], discovered new types of "parado=-=xes" [12, 15, 20, 48, 50]-=-, and proved that detecting Braess's Paradox (even in its worst-possible manifestation) is algorithmically di#cult [42]. In a related model with finitely many agents, each controlling a strictly posit... |

6 | Avoiding Paradoxes in Routing Games - Altman, Azouzi, et al. - 2001 |

5 |
The prevalence of paradoxes in transportation equilibrium problems
- Steinberg, Stone
- 1988
(Show Context)
Citation Context |

5 |
Braess’s paradox of trac flow
- Murchland
- 1970
(Show Context)
Citation Context ...21, 32, 33, 34, 35, 41, 43, 47, 49]). In the past several decades, much of the work on this tra#c model has been inspired by a "paradox" first discovered by Braess [6] and later reported by =-=Murchland [31]-=- (see also [2] for a non-technical account). The essence of Braess's Paradox is captured by the example shown in Figure 1, where the edges are labeled with their latency functions (each a function of ... |

5 |
The existence, uniqueness and stability of tra±c equilibria
- Smith
- 1979
(Show Context)
Citation Context ...: Braess's Paradox. The addition of an intuitively helpful link can negatively impact all of the users of a congested network. of and extensions to this tra#c model have been studied (see for example =-=[1, 11, 16, 21, 32, 33, 34, 35, 41, 43, 47, 49]). In the -=-past several decades, much of the work on this tra#c model has been inspired by a "paradox" first discovered by Braess [6] and later reported by Murchland [31] (see also [2] for a non-techni... |

5 | Solutions to the Optimal Network Design Problem with Shipment Related to Transportation Cost - Boyce, Soberanes - 1979 |

5 | The Counterintuitive in Conflict and Cooperation - Cohen - 1988 |

5 | Sel Trac Allocation for Server Farms - Czumaj, Krysta, et al. |

4 | The economics of tra±c congestion - Arnott, Small - 1994 |

4 | The existence of equivalent mathematical programs for certain mixed equilibrium traffic assignment problems - Bennett - 1993 |

4 | Survey of Leader-Follower Concepts in Hierarchical DecisionMaking - Cruz - 1980 |

3 |
A Generic Analysis of Selfish Routing. Working Paper, available at http://www.orie.cornell.edu/ friedman/papers.htm
- Friedman
- 2001
(Show Context)
Citation Context ...etwork flow, with the amount of flow between a pair of nodes in the network equal to the rate of traffic between the two nodes. A Nash equilibrium in the aforementioned noncooperative game 1 Friedman =-=[19]-=- has pointed out that this equilibrium arises not only in settings where network users can select their own paths, but also in network protocols that determine routes for users via a shortest path com... |

3 |
Stable dynamics transportation systems
- Nesterov, Palma
- 2000
(Show Context)
Citation Context |

2 |
The resistance of an electrical network
- Cohn
- 1950
(Show Context)
Citation Context ...roportional to its congestion; this result is implicit in the work of Dafermos and Sparrow [13], and other properties of this special case have been investigated in the context of electrical networks =-=[5, 9]-=-. Corollary 4.2 Let G be a network in which each edge latency function ℓe is of the form ℓe(x) =aex. Then for any rate vector r, aflowfeasiblefor(G, r, ℓ) is optimal if and only if it is at Nash equil... |

2 | Stable flows in transportation networks
- Nesterov
- 1999
(Show Context)
Citation Context |

2 | Stackelberg Di#erential Games in Economic Models - Bagchi - 1984 |

1 | Is Selfish Routing? 259 - Bad - 2000 |

1 | Bounding the inefficiency of Nash equilibria in nonatomic congestion games - ROUGHGARDEN, TARDOS |

1 |
How bad is selfish routing? http://www.cs.cornell.edu/timr
- Roughgarden, Tardos
(Show Context)
Citation Context ... simplicity only), we may apply the Karush-Kuhn-Tucker Theorem (see, e.g., [25]) to a convex program of the form (NLP ) to derive the following characterization of optimal flows (see the full version =-=[28]-=- for details): Lemma 2.4 A flow f is optimal for a convex program of the form (NLP ) if and only if for every i # {1, . . . , k} and P 1 , P 2 # P i with fP1 > 0, c # P1 (f) # c # P2 (f). The striking... |

1 |
Bounding the inefficiency of Nash equilibria
- Roughgarden, Tardos
- 2002
(Show Context)
Citation Context ... remark that Theorems 3.1 and 4.5 can be extended to a broader class of games (including games without the structure provided by a network); this direction of research is pursued in a companion paper =-=[46]-=-. 5.1 Flows at Approximate Nash Equilibrium It is unreasonable to expect agents to be able to evaluate the latency of different paths with arbitrary precision. We next investigate the sensitivity of o... |

1 |
Characterizing Braess's paradox for tra#c networks
- Hagstrom, Abrams
- 2001
(Show Context)
Citation Context ...w roads in congested cities [25, 31], researchers attempted to classify networks in which the addition of a single link could degrade network performance [18, 51], discovered new types of "parado=-=xes" [12, 15, 20, 48, 50]-=-, and proved that detecting Braess's Paradox (even in its worst-possible manifestation) is algorithmically di#cult [42]. In a related model with finitely many agents, each controlling a strictly posit... |

1 |
Bounding the ine#ciency of Nash equilibria
- Roughgarden, Tardos
- 2002
(Show Context)
Citation Context ... remark that Theorems 3.1 and 4.5 can be extended to a broader class of games (including games without the structure provided by a network); this direction of research is pursued in a companion paper =-=[46]-=-. 5.1 Flows at Approximate Nash Equilibrium It is unreasonable to expect agents to be able to evaluate the latency of di#erent paths with arbitrary precision. We next investigate the sensitivity of ou... |

1 | The limits of inefficiency of Nash equilibria. In preparation. [33] Y. Sheffi. Urban Transportation Networks: Equilibrium Analysis with Mathematical Programming Methods - Roughgarden, Tardos - 1985 |

1 | Properties of dynamic tra#c equilibrium involving bottlenecks, including a paradox and metering - Arnott, Palma, et al. - 1993 |

1 | Computing economic equilibria on a#ne networks with Lemke's algorithm - Asmuth, Eaves, et al. - 1979 |

1 | Secrets of network success - Bean - 1996 |

1 | Competitive routing in multicast communications - Boulogne, Altman - 2001 |

1 | The Downs-Thomson e#ect in a Markov process - Calvert - 1997 |

1 | An extended tra#c assignment model with applications to two-way tra#c - Dafermos - 1971 |