An approximate truthful mechanism for combinatorial auctions with single parameter agents
SODA '03 Proceedings of the fourteenth annual ACM-SIAM symposium on Discrete algorithms
PAC Bounds for Multi-armed Bandit and Markov Decision Processes
COLT '02 Proceedings of the 15th Annual Conference on Computational Learning Theory
Multi-unit auctions with budget-constrained bidders
Proceedings of the 6th ACM conference on Electronic commerce
Online ascending auctions for gradually expiring items
SODA '05 Proceedings of the sixteenth annual ACM-SIAM symposium on Discrete algorithms
AdWords and Generalized On-line Matching
FOCS '05 Proceedings of the 46th Annual IEEE Symposium on Foundations of Computer Science
Anytime algorithms for multi-armed bandit problems
SODA '06 Proceedings of the seventeenth annual ACM-SIAM symposium on Discrete algorithm
Truthful auctions for pricing search keywords
EC '06 Proceedings of the 7th ACM conference on Electronic commerce
An incentive-compatible multi-armed bandit mechanism
Proceedings of the twenty-sixth annual ACM symposium on Principles of distributed computing
Sponsored Search Auctions with Reserve Prices: Going Beyond Separability
WINE '08 Proceedings of the 4th International Workshop on Internet and Network Economics
Adaptive Incentive-Compatible Sponsored Search Auction
SOFSEM '09 Proceedings of the 35th Conference on Current Trends in Theory and Practice of Computer Science
A truthful learning mechanism for contextual multi-slot sponsored search auctions with externalities
Proceedings of the 13th ACM Conference on Electronic Commerce
| Hi-index | 0.00 |
This paper presents an online sponsored search auction that motivates advertisers to report their true budget, arrival time, departure time, and value per click. The auction is based on a modified Multi-Armed Bandit (MAB) mechanism that allows for advertisers who arrive and depart in an online fashion, have a value per click, and are budget constrained. In tackling the problem of truthful budget, arrival and departure times, it turns out that it is not possible to achieve truthfulness in the classical sense (which we show in a companion paper). As such, we define a new concept called δ-gain. δ-gain bounds the utility a player can gain by lying as opposed to his utility when telling the truth. Building on the δ-gain concept we define another new concept called relative Ɛ-gain, which bounds the relative ratio of the gain a player can achieve by lying with respect to his true utility. We argue that for many practical applications if the δ-gain and or the relative Ɛ-gain are small, then players will not invest time and effort in making strategic choices but will truthtell as a default strategy. These concepts capture the essence of dominant strategy mechanisms as they lead the advertiser to choose truthtelling over other strategies. In order to achieve δ-gain truthful mechanism this paper also presents a new payment scheme, Time series Truthful Payment Scheme (TTPS), for an online budget-constrained auction mechanism. The payment scheme is a generalization of the VCG principles for an online scheduling environment with budgeted players. Using the concepts of δ-gain truthful we present the only known budget-constrained sponsored search auction with truthful guarantees on budget, arrivals, departures, and valuations. Previous works that deal with advertiser budgets only deal with the non-strategic case.