Abstract
We study how collusion affects the social cost in atomic splittable routing games. Suppose that players form coalitions and each coalition behaves as if it were a single player controlling all the flows of its participants. We investigate the following question: under what conditions would the social cost of the post-collusion equilibrium be bounded by the social cost of the pre-collusion equilibrium?
We show that if (i) the network is “well-designed” (satisfying a natural condition), and (ii) the delay functions are affine, then collusion is always beneficial for the social cost in the equilibrium flows. On the other hand, if either of the above conditions is unsatisfied, collusion can worsen the social cost.
Our main technique is a novel flow-augmenting algorithm to build equilibrium flows. Our positive result for collusion is obtained by applying this algorithm simultaneously to two different flow value profiles of players and observing the difference in the derivatives of their social costs. Moreover, for a non-trivial subclass of selfish routing games, this algorithm finds the exact equilibrium flows in polynomial time.
Similar content being viewed by others
Notes
Even though the result for affine delays as stated in [9] is only for series-parallel networks, the technique of [9] indeed also works for well-designed networks. This is because a well-designed network always satisfies the nesting property (see Sect. 3.3 for definition). A minor contribution of this paper is to identify a larger class of networks for this result of [9] to hold.
The proof proceeds as follows. Let σ′ be the profile of k players, each with the same amount of flow. Let σ be the profile of the post-collusion equilibrium flow. Observe that σ dominates σ′. It is established in [16] that the social cost of the equilibrium flow for σ′ has social cost no larger than the pre-collusion Wardrop equilibrium. Now applying Lemma 5 to compare social costs of the equilibria flows for σ′ and for σ would give the proof.
This follows from the fact that in an equilibrium flow of symmetric players, the flows of all players are identical [16].
To simplify our presentation, we slightly abuse notation by writing a value that goes to infinity as a fixed value t x .
Since we increase the flow by an infinitesimal amount ϵ, this algorithm, along with the main algorithm in the next section, do not run in polynomial time. But they can be easily modified to run in polynomial time. We choose to present in this manner since we consider it as simpler and it makes the analysis in Sect. 3.4 go smoother.
To be precise, C(σ,t) is not differentiable in the breakpoints t i (σ). Hence the domain of \(\frac{d C(\sigma,t)}{d t}\) is [0,1]∖{t i (σ)}∀i . For our purpose, it suffices to consider the open intervals between these breakpoints, because the functions C(σ,t) and C(σ′,t) are both continuous.
Recall our earlier remark at the end of Sect. 3.1: the growing speed of the social cost using a larger set of paths \(\mathcal {P}_{h(\sigma, t^{*})}\), which is \(Y_{h(\sigma,t^{*})} t^{*} + Z_{h(\sigma,t^{*})}\), is less than the growing speed of the social cost using a smaller set of paths \(\mathcal {P}_{h(\sigma', t^{*})}\), which is \(Y_{h(\sigma',t^{*})} t^{*} + Z_{h(\sigma',t^{*})}\), under a certain condition, which is \(t^{*} > t_{h(\sigma,t^{*})-1}\) and it is proved in Lemma 22.
References
Ahuja, R., Magnanti, T., Orlin, J.: Network Flows. Prentice Hall, New York (1993)
Aland, S., Dumrauf, D., Gairing, M., Monien, B., Schoppmann, F.: Exact price of anarchy for polynomial congestion games. In: STACS, pp. 218–229 (2006)
Albers, S.: Strong and pareto price of anarchy in congestion games. SIAM J. Comput. 38(6), 2273–2302 (2009)
Altman, E., Basar, T., Jimenez, T., Shimkin, N.: Competitive routing in networks with polynomial costs. IEEE Trans. Autom. Control 47, 92–96 (2002)
Andelman, N., Feldman, M., Mansour, Y.: Strong price of anarchy. In: SODA, pp. 189–198 (2007)
Aumann, R.: Acceptable points in general cooperative n-person games. Contrib. Theory Games 4 (1959)
Awerbuch, B., Azar, Y., Epstein, A.: The price of routing unsplittable flow. In: STOC, pp. 57–66 (2005)
Bhaskar, U., Fleischer, L., Hoy, D., Huang, C.-C.: Equilibria of atomic flow games are not unique. In: SODA, pp. 748–757 (2009)
Bhaskar, U., Fleischer, L., Huang, C.-C.: The price of collusion in series-parallel networks. In: IPCO, pp. 313–326 (2010)
Bollobás, B.: Modern Graph Theory. Springer, Berlin (1998)
Bonifaci, V., Harks, T., Schäfer, G.: Stackelberg routing in arbitrary networks. Math. Oper. Res. 35(2), 330–346 (2010)
Catoni, S., Pallottino, S.: Traffic equilibrium paradoxes. Transp. Sci. 25(3), 240–244 (1991)
Chien, S., Sinclair, A.: Strong and pareto price of anarchy in congestion games. In: ICALP, pp. 279–291 (2009)
Christodoulou, G., Koutsoupias, E.: The price of anarchy of finite congestion games. In: STOC, pp. 67–73 (2005)
Christodoulou, G., Koutsoupias, E., Spirakis, P.G.: On the performance of approximate equilibria in congestion games. Algorithmica 61(2), 116–140 (2011)
Cominetti, R., Correa, J.R., Stier-Moses, N.E.: The impact of oligopolistic competition in networks. Oper. Res. 57(6), 1421–1437 (2009)
Correa, J.R., Schulz, A.S., Stier-Moses, N.E.: Selfish routing in capacitated networks. Math. Oper. Res. 29(4), 961–976 (2004)
Duffin, R.: Topology of series-parallel networks. J. Math. Anal. Appl. 10(2), 303–318 (1965)
Epstein, A., Feldman, M., Mansour, Y.: Strong equilibrium in cost-sharing connection games. In: EC, pp. 84–92 (2007)
Fiat, A., Levy, M., Kaplan, H., Olonetsky, S.: Strong price of anarchy for machine load balancing. In: ICALP, pp. 583–594 (2007)
Fleischer, L.: Linear tolls suffice: new bounds and algorithms for tolls in single source networks. Theor. Comput. Sci. 348(2–3), 217–225 (2005)
Fleischer, L., Jain, K., Mahdian, M.: Tolls for heterogeneous selfish users in multicommodity networks and generalized congestion games. In: FOCS, pp. 277–285 (2004)
Fotakis, D., Kontogiannis, S., Koutsoupias, E., Mavronicolas, M., Spirakis, P.: The structure and complexity of Nash equilibria for a selfish routing game. In: ICALP, pp. 123–134 (2002)
Fotakis, D., Kontogiannis, S., Spirakis, P.: Symmetry in network congestion games: pure equilibria and anarchy cost. In: WAOA, pp. 161–175 (2005)
Fotakis, D., Kontogiannis, S.C., Spirakis, P.: Selfish unsplittable flows. Theor. Comput. Sci. 226–239 (2005)
Fotakis, D., Kontogiannis, S., Spirakis, P.: Atomic congestion games among coalitions. ACM Trans. Algorithms 4, 4 (2008). The conference version appeared in ICALP 2006
Hariharan, R., Kavitha, T., Mehlhorn, K.: Faster algorithms for minimum cycle basis in directed graphs. SIAM J. Comput. 38(4), 1430–1447 (2008)
Harks, T.: Stackelberg strategies and collusion in network games with splittable flow. Theory Comput. Syst. 781–802 (2011)
Harks, T., Klimm, M., Möhring, R.: Strong Nash equilibria in games with the lexicographical improvement property. In: WINE, pp. 463–470 (2009)
Harks, T., Höfer, M., Klimm, M., Skopalik, A.: Computing pure Nash and strong equilibria in bottleneck congestion games. In: ESA, pp. 29–38 (2010)
Hayrapetyan, A., Tardos, E., Wexler, T.: The effect of collusion in congestion games. In: STOC, pp. 89–98 (2006)
Karakostas, G., Kolliopoulos, S.G.: Edge pricing of multicommodity networks for heterogeneous selfish users. In: FOCS, pp. 268–276 (2004)
Koutsoupias, E., Papadimitriou, C.: Worst-case equilibria. In: STACS, pp. 404–413 (1999)
Lin, H., Roughgarden, T., Tardos, E., Walkover, A.: Stronger bounds on Braess’s paradox and the maximum latency of selfish routing. SIAM J. Discrete Math. 25(4), 1667–1686 (2011)
Minoux, M.: A polynomial algorithm for minimum quadratic cost flow problems. Eur. J. Oper. Res. 18, 377–387 (1984)
Nocedal, J., Wright, S.: Numerical Optimization. Springer, Berlin (2006)
Orda, A., Rom, R., Shimkin, N.: Competitive routing in multiuser communication networks. IEEE/ACM Trans. Netw. 1(5), 510–521 (1993)
Rosen, J.B.: Existence and uniqueness of equilibrium points for concave n-person games. Econometrica 33(3), 520–534 (1965)
Roughgarden, T.: Stackelberg scheduling strategies. SIAM J. Comput. 33(2), 332–350 (2004)
Roughgarden, T.: Selfish Routing and the Price of Anarchy. MIT Press, Cambridge (2005)
Roughgarden, T.: On the severity of Braess’s paradox: designing networks for selfish users is hard. J. Comput. Syst. Sci. 72(5), 922–953 (2006)
Roughgarden, T.: Intrinsic robustness of the price of anarchy. In: STOC, pp. 513–522 (2009)
Roughgarden, T., Schoppmann, F.: Local smoothness and the price of anarchy in atomic splittable congestion game. In: SODA (2011)
Roughgarden, T., Tardos, E.: How bad is selfish routing? J. ACM 49(2), 236–259 (2002)
Swamy, C.: The effectiveness of Stackelberg strategies and tolls for network congestion games. In: SODA, pp. 1133–1142 (2007)
Valiant, G., Roughgarden, T.: Braess’s paradox in large random graphs. Random Struct. Algorithms 37(4), 495–515 (2010)
Wan, C.: Coalitions in nonatomic network congestion games. arXiv:1203.5822v1
Wardrop, J.G.: Some theoretical aspects of road traffic research. In: Proc. Institute of Civil Engineers, Pt. II, vol. 1, pp. 325–378 (1952)
Acknowledgements
I thank Lisa Fleischer and Kurt Mehlhorn for their comments on earlier drafts and Khaled Elbassioni and Michael Sagraloff for discussions on several technical issues. I also thank the two anonymous reviewers for their comments that help improve the presentation of this paper. Research supported by an Alexander von Humboldt fellowship.
Author information
Authors and Affiliations
Corresponding author
Appendix: Proof of Proposition 2
Appendix: Proof of Proposition 2
In the following proof, we implicitly assume that all edges are directed from the origin to the destination and the graph is acyclic in the directed edges. This does not hurt generality, since any edge not on an o-d path is not used in an optimal flow, and no optimal flow is sent along a directed cycle.
Proof of Proposition 2
Let O(t) and O(t′) be optimal flows of values t and t′ and t>t′. Suppose that the proposition does not hold. Then in the pseudo-flow z=O(t)−O(t′), there exists some edge e ∗ on which \(z_{e^{*}}<0\). Let G′=(V′,E′) be the largest subgraph of G containing e such that G′ is also series-parallel and ∑ e∈E′ z e <0. Then there exists another subgraph G″=(V″,E″) such that ∑ e∈E″ z e >0 and G″ and G′ share the same two terminals o′ and d′.
Claim
There exists an o′-d′ path p′ in G′ so that z e <0, ∀e∈p′ and another o′-d′ path p″ in G″ so that z e >0, ∀e∈p″.
If the claim holds, then ∑ e∈p′ O e (t)<∑ e∈p′ O e (t′) and ∑ e∈p″ O e (t)>∑ e∈p″ O e (t′). Recall that optimal flow is an equilibrium flow of one player, so we can apply Lemma 8 and the preceding two inequalities to derive
a contradiction. □
Proof of the Claim
We only prove the first part. Suppose that G′ is composition of \(G'_{1}=(V'_{1},E'_{1})\) and \(G'_{2}=(V'_{2},E'_{2}) \). If G′ is resulted from the series-composition of \(G'_{1}\) and \(G'_{2}\), then \(\sum_{e\in E'_{1}} z_{e} = \sum_{e\in E'_{2}} z_{e}<0\). If G′ is resulted from the parallel-composition of \(G'_{1}\) and \(G'_{2}\), then either \(\sum_{e\in E'_{1}} z_{e} <0\) or \(\sum_{e\in E'_{2}} z_{e}<0\). In both cases, we can apply induction to find out a path p′ in \(G'_{1}\) and/or \(G'_{2}\) so that z e <0 if e∈p′. □
Rights and permissions
About this article
Cite this article
Huang, CC. Collusion in Atomic Splittable Routing Games. Theory Comput Syst 52, 763–801 (2013). https://doi.org/10.1007/s00224-012-9421-4
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00224-012-9421-4