Approximation Algorithms for Secondary Spectrum Auctions

We study combinatorial auctions for the secondary spectrum market. In this market, short-term licenses shall be given to wireless nodes for communication in their local neighborhood. In contrast to the primary market, channels can be assigned to multiple bidders, provided that the corresponding devices are well separated such that the interference is sufficiently low. Interference conflicts are described in terms of a conflict graph in which the nodes represent the bidders and the edges represent conflicts such that the feasible allocations for a channel correspond to the independent sets in the conflict graph. In this paper, we suggest a novel LP formulation for combinatorial auctions with conflict graph using a non-standard graph parameter, the so-called inductive independence number. Taking into account this parameter enables us to bypass the well-known lower bound of \Omega(n^{1-\epsilon}) on the approximability of independent set in general graphs with n nodes (bidders). We achieve significantly better approximation results by showing that interference constraints for wireless networks yield conflict graphs with bounded inductive independence number. Our framework covers various established models of wireless communication, e.g., the protocol or the physical model. For the protocol model, we achieve an O(\sqrt{k})-approximation, where k is the number of available channels. For the more realistic physical model, we achieve an O(\sqrt{k} \log^2 n) approximation based on edge-weighted conflict graphs. Combining our approach with the the LP-based framework of Lavi and Swamy, we obtain incentive compatible mechanisms for general bidders with arbitrary valuations on bundles of channels specified in terms of demand oracles.

💡 Research Summary

The paper tackles the combinatorial auction problem that arises in secondary spectrum markets, where a limited number of wireless channels must be allocated to multiple bidders under interference constraints. Unlike primary markets, a single channel can be reused by several users as long as the devices are sufficiently separated, which is naturally modeled by a conflict graph: vertices represent bidders and edges denote pairs that cannot simultaneously use the same channel. A feasible allocation for a channel therefore corresponds to an independent set in this graph.

A major obstacle in designing approximation algorithms for such settings is the well‑known hardness of approximating the maximum independent set in general graphs, which admits only an Ω(n^{1‑ε}) lower bound. To circumvent this barrier, the authors introduce a non‑standard graph parameter called the inductive independence number (ρ). ρ is defined with respect to an ordering of the vertices; when vertices are processed in that order, each vertex can be adjacent to at most ρ previously processed vertices that belong to any independent set. In many wireless interference models—both the protocol model (distance‑based) and the physical (SINR) model—this number is provably small (constant or poly‑logarithmic), because the geometric nature of interference limits how many already‑selected nodes can conflict with a new one.

Leveraging ρ, the authors formulate a novel linear program (LP). For each bidder i and each bundle S of channels, a variable x_{i,S} denotes the fractional probability that i receives exactly S. The objective maximizes total expected valuation (social welfare). The constraints enforce (1) that each bidder receives at most one bundle, (2) that for each channel the set of bidders assigned to it respects the conflict graph, and crucially (3) that the conflict constraints are relaxed by a factor of ρ, i.e., the sum of x‑variables over any set of mutually conflicting bidders is bounded by ρ. This LP can be solved in polynomial time and provides an upper bound on the optimal integral solution.

The rounding scheme is built around the inductive ordering that defines ρ. Vertices are examined sequentially; when a vertex is considered, it is added to the solution with probability proportional to its fractional value divided by ρ, provided that none of its already‑selected neighbors conflict on the same channel. This simple randomized rounding yields an expected welfare of at least 1/ρ times the LP optimum, establishing a ρ‑approximation.

Applying this framework to specific wireless models yields concrete approximation ratios. In the protocol model, where interference is captured by unit‑disk graphs, ρ can be bounded by O(√k) where k is the number of available channels. Consequently the algorithm achieves an O(√k)‑approximation. In the more realistic physical (SINR) model, interference is represented by edge‑weighted conflict graphs; the authors show that ρ = O(√k·log n). After a careful logarithmic scaling of the rounding probabilities, the final guarantee becomes O(√k·log² n). Both results dramatically improve over the generic independent‑set hardness and are the first to provide sub‑linear (in k) approximation factors for multi‑channel spectrum auctions.

Beyond approximation, the paper integrates the LP‑based approach with the mechanism design framework of Lavi and Swamy. By assuming that each bidder can answer a demand oracle (i.e., given a set of prices, return a utility‑maximizing bundle), the authors construct a truthful (in expectation) auction mechanism. The LP solution is used to compute fractional allocations and VCG‑style payments; the randomized rounding preserves truthfulness because the allocation rule is a linear function of the LP variables, and the payments are calibrated to the expected marginal contribution of each bidder. This yields an incentive‑compatible mechanism for bidders with arbitrary combinatorial valuations over bundles of channels.

The paper’s contributions can be summarized as follows:

1. Introduction of the inductive independence number ρ as a tool to capture the limited interference structure of wireless networks.
2. A unified LP formulation that incorporates ρ‑scaled conflict constraints, enabling polynomial‑time solvability.
3. A simple, order‑based randomized rounding that guarantees a 1/ρ approximation.
4. Concrete bounds of O(√k) for the protocol model and O(√k·log² n) for the physical model, both substantially better than the Ω(n^{1‑ε}) barrier.
5. Extension to truthful auction mechanisms via the Lavi‑Swamy framework, requiring only demand oracles from bidders.

Overall, the work demonstrates that exploiting geometric and physical properties of wireless interference can break longstanding hardness barriers in combinatorial auctions, delivering both strong approximation guarantees and practical, incentive‑compatible market mechanisms.