Threshold mechanisms for dynamic procurement with abandonment

From MaRDI portal
Publication:6546307





This paper presents a model of a two-round dynamic procurement auction with abandonment, where the agents leave the auction if they do not meet their overhead costs in a given round. The authors focus on threshold mechanisms, as they are widely used in practice, and show that they are ex-post incentive compatible for the dynamic auction model. The thresholds are similar to setting reserves for revenue maximization in regular auctions. Next, the optimization problem for finding the optimal set of thresholds is studied. It is shown that if the distribution \(F\) for overhead costs is regular (as defined later), the optimal thresholds are independent of the number of agents participating in the auction. It is proven that if the underlying distribution \(F\) satisfies certain properties, the optimal thresholds will be monotone, meaning that the optimal threshold for saving \(i\) agents is lower than the optimal threshold for saving \(j\) agents for any \(i>j\). Moreover, this monotonicity helps divide the optimization problem into \(n\) separate problems, which ultimately leads to an efficient algorithm to calculate the optimal thresholds in parallel.\N\NFor the entire collection see [Zbl 1537.91004].











This page was built for publication: Threshold mechanisms for dynamic procurement with abandonment

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6546307)