Conference proceeding
Threshold Mechanisms for Dynamic Procurement with Abandonment
ALGORITHMIC GAME THEORY, SAGT 2023, v 14238, pp 383-400
01 Jan 2023
Featured in Collection : UN Sustainable Development Goals @ Drexel
Abstract
We study a dynamic model of procurement auctions in which the agents (sellers) will abandon the auction if their utility does not satisfy their private target, in any given round. We call this "abandonment" and analyze its consequences on the overall cost to the mechanism designer (buyer), as it reduces competition in future rounds of the auction and drives up the price. We show that in order to maintain competition and minimize the overall cost, the mechanism designer has to adopt an inefficient (per-round) allocation, namely to assign the demand to multiple agents in a single round. We focus on threshold mechanisms as a simple way to achieve ex-post incentive compatibility, akin to reserves in revenue-maximizing forward auctions. We then consider the optimization problem of finding the optimal thresholds. We show that even though our objective function does not have the optimal substructure property in general, if the underlying distributions satisfy some regularity properties, the global optimal solution lies within a region where the optimal thresholds are monotone and can be calculated with a greedy approach, or even more simply in a parallel fashion.
Metrics
7 Record Views
Details
- Title
- Threshold Mechanisms for Dynamic Procurement with Abandonment
- Creators
- Ali Khodabakhsh - The University of Texas at AustinEvdokia Nikolova - The University of Texas at AustinEmmanouil Pountourakis - Drexel UniversityJimmy Horn - Wind
- Contributors
- A Deligkas (Editor)A Filos-Ratsikas (Editor)
- Publication Details
- ALGORITHMIC GAME THEORY, SAGT 2023, v 14238, pp 383-400
- Series
- Lecture Notes in Computer Science
- Publisher
- Springer Nature
- Number of pages
- 18
- Grant note
- CCF-2218813 / NSF; National Science Foundation (NSF)
- Resource Type
- Conference proceeding
- Language
- English
- Academic Unit
- Computer Science (Computing)
- Web of Science ID
- WOS:001156199000022
- Scopus ID
- 2-s2.0-85172147341
- Other Identifier
- 991021869109304721
UN Sustainable Development Goals (SDGs)
This publication has contributed to the advancement of the following goals:
InCites Highlights
Data related to this publication, from InCites Benchmarking & Analytics tool:
- Collaboration types
- Domestic collaboration
- Web of Science research areas
- Computer Science, Theory & Methods
- Mathematics, Applied