Dynamics of multidimensional Simple Clock Auctions

Simple Clock Auctions (SCA) are a mechanism commonly used in spectrum auctions to sell lots of frequency bandwidths. We study such an auction with one player having access to perfect information against straightforward bidders. When the opponents’ valuations satisfy the ordinary substitutes condition, we show that it is optimal to bid on a fixed lot overtime. In this setting, we consider a continuous-time version of the SCA auction in which the prices follow a differential inclusion with a piecewise-constant dynamics. We show that there exists a unique solution in the sense of Filippov. This guarantees that the continuous-time model coincides with the limit of the discrete-time auction when price increments tend to zero. Moreover, we show that the value function of this limit auction is piecewise linear (though possibly discontinuous). Finally, we illustrate these results by analyzing a simplified version of the multiband Australian spectrum auction of 2017.

💡 Research Summary

This paper investigates the dynamics and optimal strategies of multidimensional Simple Clock Auctions (SCA), a mechanism widely used in spectrum license sales. The authors consider a setting with two players: one “informed” player who knows the exact valuation functions of all opponents, and a second player who follows the Straightforward Bidding (SB) strategy, i.e., maximizing immediate profit given the current price without anticipating future price changes.

The first major contribution is a structural result for the discrete‑time auction. Assuming the opponent’s valuation satisfies the ordinary substitutes condition—a standard combinatorial auction property stating that raising the price of a good never reduces demand for any other good—the authors prove (Theorem 5) that a constant bundle (a fixed demand vector) is optimal for the informed player, regardless of the number of item categories M. This extends earlier one‑dimensional results to the full multidimensional case and dramatically simplifies the optimal control problem.

To connect the discrete model with a continuous‑time limit, the paper introduces a differential inclusion describing price evolution as the price increment vector ε tends to zero. The inclusion has the form
 dp/dt ∈ F(p)
where F(p) is a piecewise‑constant, polyhedral set‑valued map. The discontinuities of F correspond to the “tropical hypersurfaces” generated by the opponent’s valuation; each cell of the tropical complex represents a region where the opponent’s demand is constant. By invoking Filippov’s theory of differential inclusions and constructing a sup‑norm Lyapunov function, the authors establish both existence and uniqueness of Filippov solutions (Theorems 6 and 10). This is notable because the vector field has multiple discontinuity surfaces, a situation where uniqueness is generally undecidable.

The paper then studies the value function of the continuous‑time limit. It shows (Theorem 12) that the value is piecewise linear in the price vector, though it may be discontinuous at points where the price trajectory crosses a tropical hypersurface. Moreover, the authors prove (Theorem 13) that this continuous‑time value coincides with the limit of the discrete‑time value functions as ε → 0, thereby confirming that the continuous model faithfully captures the asymptotic behavior of the original auction.

A concrete illustration is provided using a simplified version of the 2017 Australian multi‑band spectrum auction. The model is reduced to two operators and two frequency bands. By specifying concrete valuation functions for both parties, the authors compute the optimal constant bid for the informed player and the resulting price dynamics. The example confirms that the informed player’s optimal strategy is indeed constant, and that price adjustments occur precisely when the opponent’s demand switches across the tropical hypersurface.

Overall, the paper makes four interrelated contributions: (1) it identifies ordinary substitutes as the key condition guaranteeing the optimality of a constant bidding strategy in multidimensional SCAs; (2) it formulates a rigorous continuous‑time limit via a differential inclusion; (3) it leverages Filippov’s framework and tropical geometry to prove existence, uniqueness, and piecewise‑linear value structure; and (4) it validates the theory on a realistic spectrum‑auction case study. These results deepen the theoretical understanding of clock auctions, provide a tractable method for computing optimal strategies against straightforward bidders, and open avenues for applying tropical‑geometric tools to other auction‑dynamic problems.