Enforcing efficient equilibria in network design games via subsidies

The efficient design of networks has been an important engineering task that involves challenging combinatorial optimization problems. Typically, a network designer has to select among several alternatives which links to establish so that the resulting network satisfies a given set of connectivity requirements and the cost of establishing the network links is as low as possible. The Minimum Spanning Tree problem, which is well-understood, is a nice example. In this paper, we consider the natural scenario in which the connectivity requirements are posed by selfish users who have agreed to share the cost of the network to be established according to a well-defined rule. The design proposed by the network designer should now be consistent not only with the connectivity requirements but also with the selfishness of the users. Essentially, the users are players in a so-called network design game and the network designer has to propose a design that is an equilibrium for this game. As it is usually the case when selfishness comes into play, such equilibria may be suboptimal. In this paper, we consider the following question: can the network designer enforce particular designs as equilibria or guarantee that efficient designs are consistent with users’ selfishness by appropriately subsidizing some of the network links? In an attempt to understand this question, we formulate corresponding optimization problems and present positive and negative results.

💡 Research Summary

The paper studies network design games in which selfish users share the cost of edges along their chosen paths. Because each player minimizes her own cost, the resulting Nash equilibria can be far from socially optimal, leading to a high price of stability. To mitigate this inefficiency, the authors introduce subsidies: a central authority may partially pay for selected edges, thereby influencing players’ incentives. Two optimization problems are defined. Stable Network Enforcement (SNE) asks, given a target network T, for the minimum total subsidy that makes T a Nash equilibrium. Stable Network Design (SND) asks, given a budget, to construct a network and a subsidy allocation that yields an equilibrium while minimizing the total social cost.

The authors first show that SNE can be expressed as a linear program with an exponential number of constraints; using the ellipsoid method it can be solved in polynomial time. For broadcast games (one player per non‑root node, each needing a path to a common root) the LP simplifies dramatically to a formulation with O(n) variables and O(n²) constraints, making SNE efficiently solvable in practice.

In contrast, SND is shown to be NP‑hard even for broadcast instances. Detecting whether the minimum spanning tree (MST) can be enforced as an equilibrium without any subsidies is already NP‑hard, implying that deciding whether the price of stability equals 1 is computationally intractable. Moreover, approximating the price of stability for broadcast games is APX‑hard, extending known hardness results for computing equilibria in network design games.

The paper then quantifies how much subsidy is needed to enforce an MST as an equilibrium. It proves an upper bound of 37 % of the MST’s total weight: a carefully constructed subsidy scheme guarantees that no player can improve by deviating. This bound is tight, as there exist instances where at least 37 % of the MST weight is necessary. For the all‑or‑nothing version of SNE (each edge is either fully subsidized or not at all), the required subsidy can be as high as 61 % of the MST weight, and the problem becomes hard to approximate within any constant factor.

Overall, the work establishes a clear dichotomy: while enforcing a given network can be done efficiently via linear programming, designing a low‑cost network under a subsidy budget is computationally hard, and substantial subsidies may be required to achieve efficiency. The results illuminate the trade‑off between budget constraints and equilibrium efficiency in strategic network formation, and they open several avenues for future research on approximation algorithms and alternative incentive mechanisms.