AdWords and generalized on-line matching

@article{Mehta2005AdWordsAG,
  title={AdWords and generalized on-line matching},
  author={Aranyak Mehta and Amin Saberi and Umesh V. Vazirani and Vijay V. Vazirani},
  journal={46th Annual IEEE Symposium on Foundations of Computer Science (FOCS'05)},
  year={2005},
  pages={264-273}
}
How does a search engine company decide what ads to display with each query so as to maximize its revenue? This turns out to be a generalization of the online bipartite matching problem. We introduce the notion of a tradeoff revealing LP and use it to derive two optimal algorithms achieving competitive ratios of 1-1/e for this problem. 

Figures from this paper

Randomized Online Algorithms for Adwords
TLDR
A randomized online algorithm achieving a competitive ratio of (k-TYPICAL, k ∈ Z+,) is given, which seems a reasonable assumption for a “typical” instance, at least for moderate values of k.
Online Matching with High Probability
TLDR
A tail inequality bound is given, namely that Ranking is (1 − 1 e − α)-competitive with probability at least 1 − e n where n is the size of the maximum matching in the instance.
Two-sided Online Bipartite Matching and Vertex Cover: Beating the Greedy Algorithm
We consider the generalizations of two classical problems, online bipartite matching and ski rental, in the field of online algorithms, and establish a novel connection between them.
Online Bipartite Matching with Reusable Resources
TLDR
This work gives the first approximation factor beating 0.5, namely a 0.505 approximation, by suitably adapting and interpreting the powerful technique of Online Correlated Selection.
Online Bipartite Matching and Adwords
TLDR
This paper has stated the result for SMALL, namely an optimal, online, budgetoblivious algorithm, as a Conditional Theorem, after assuming the no-surpassing property, and believes that in typical instances, this property fails on very few queries, if any.
Online budgeted matching in random input models with applications to Adwords
TLDR
An online assignment problem, motivated by Adwords Allocation, in which queries are to be assigned to bidders with budget constraints is studied, with a tight analysis of Greedy in this model showing that it has a competitive ratio of 1 - 1/e for maximizing the value of the assignment.
Algorithms for Online Matching, Assortment, and Pricing with Tight Weight-dependent Competitive Ratios
TLDR
This paper analyzes a general problem of allocating finite inventories to heterogeneous customers arriving sequentially under the framework of competitive analysis, and develops algorithms which achieve the best-possible weight-dependent competitive ratios, which depend on the sets of feasible prices given in advance.
An Improved Analysis for a Greedy Remote-Clique Algorithm Using Factor-Revealing LPs
TLDR
The technique of factor-revealing linear programs is used to show that the greedy algorithm for the remote-clique problem actually achieves an approximation ratio of 2, which is tight.
Online Matching in Regular Bipartite Graphs with Randomized Adversary
TLDR
The competitive ratio for 2-regular graphs for any number of random decisions made by the adversary is determined and upper and lower bounds for other values of the degree are given.
Algorithms for Online Matching, Assortment, and Pricing with Tight Weight-dependent Competitive Ratios
TLDR
This paper studies a general problem of allocating finite inventories to heterogeneous customers arriving sequentially under the framework of competitive analysis, and develops algorithms which achieve the best-possible weight-dependent competitive ratios, which depend on the sets of feasible prices given in advance.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 54 REFERENCES
AdWords and Generalized On-line Matching
TLDR
The notion of a tradeoff revealing LP is introduced and used to derive two optimal algorithms achieving competitive ratios of 1-1/e for this problem of online bipartite matching.
An optimal deterministic algorithm for online b-matching
Greedy facility location algorithms analyzed using dual fitting with factor-revealing LP
TLDR
The method of dual fitting and the idea of factor-revealing LP are formalized and used to design and analyze two greedy algorithms for the metric uncapacitated facility location problem.
An optimal algorithm for on-line bipartite matching
TLDR
This work applies the general approach to data structures, bin packing, graph coloring, and graph coloring to bipartite matching and shows that a simple randomized on-line algorithm achieves the best possible performance.
A Greedy Facility Location Algorithm Analyzed Using Dual Fitting
TLDR
A natural greedyalgorithm for the metric uncapacitated facilitylo cation problem is presented and the method of dual fitting is used to analyze its approximation ratio, which turns out to be 1.861.
On-line Network Optimization Problems
There has been a reasonable amount of work on on-line network optimization in the monotone model. Usually one first analyzes the greedy algorithm that lets Hi+1 be the cheapest extension of Hi. Some
Truthful auctions for pricing search keywords
We present a truthful auction for pricing advertising slots on a web-page assuming that advertisements for different merchants must be ranked in decreasing order of their (weighted) bids. This
Multi-Item Auctions
A collection of items is to be distributed among several bidders, and each bidder is to receive at most one item. Assuming that the bidders place some monetary value on each of the items, it has been
Allocating online advertisement space with unreliable estimates
TLDR
The problem of optimally allocating online advertisement space to budget-constrained advertisers is studied and an algorithm that takes advantage of the given estimates of the frequencies of keywords to compute a near optimal solution when the estimates are accurate, while at the same time maintaining a good worst-case competitive ratio.
Auctions with Budget Constraints
TLDR
This paper presents exact and approximate algorithms for auctions with budget constraints, and presents a randomized algorithm with an approximation ratio of \(\frac{e}{e-1}\cong\) 1.582, which can be derandomized.
...
1
2
3
4
5
...