Orestis Telelis

User recruitment for mobile crowdsensing over opportunistic networks

We look into the realization of mobile crowdsensing campaigns that draw on the opportunistic networking paradigm, as practised in delay-tolerant networks but also in the emerging device-to-device communication mode in cellular networks. In particular, we ask how mobile users can be optimally selected in order to generate the required space-time paths across the network for collecting data from a set of fixed locations. The users hold different roles in these paths, from collecting data with their sensing-enabled devices to relaying them across the network and uploading them to data collection points with Internet connectivity. We first consider scenarios with deterministic node mobility and formulate the selection of users as a minimum-cost set cover problem with a submodular objective function. We then generalize to more realistic settings with uncertainty about the user mobility. A methodology is devised for translating the statistics of individual user mobility to statistics of spacetime path formation and feeding them to the set cover problem formulation. We describe practical greedy heuristics for the resulting NP-hard problems and compute their approximation ratios. Our experimentation with real mobility datasets (a) illustrates the multiple tradeoffs between the campaign cost and duration, the bound on the hopcount of space-time paths, and the number of collection points; and (b) provides evidence that in realistic problem instances the heuristics perform much better than what their pessimistic worst-case bounds suggest.

Mechanisms for multi-unit combinatorial auctions with a few distinct goods

We design and analyze deterministic truthful approximation mechanisms for multi-unit Combinatorial Auctions with only a constant number of distinct goods, each in arbitrary limited supply. Prospective buyers (bidders) have preferences over multisets of items, i.e. for more than one unit per distinct good. Our objective is to determine allocations of multisets that maximize the Social Welfare. Despite the recent theoretical advances on the design of truthful Combinatorial Auctions (for several distinct goods) and multi-unit auctions (for a single good), results for the combined setting are much scarser. Our main results are for multi-minded and submodular bidders. In the first setting each bidder has a positive value for being allocated one multiset from a prespecified demand set of alternatives. In the second setting each bidder is associated to a submodular valuation function that defines his value for the multiset he is allocated. For multi-minded bidders we design a truthful FPTAS that fully optimizes the Social Welfare, while violating the supply constraints on goods within factor (1+e) for any fixed e > 0 (i.e., the approximation applies to the constraints and not to the Social Welfare). This result is best possible, in that full optimization is impossible without violating the supply constraints. It also improves significantly upon a related result of Grandoni et al. [SODA 2010]. For submodular bidders we extend a general technique by Dobzinski and Nisan [JAIR, 2010] for multi-unit auctions, to the case of multiple distinct goods. We use this extension to obtain a PTAS that approximates the optimum Social Welfare within factor (1+e) for any fixed e > 0, without violating the supply constraints. This result is best possible as well. Our allocation algorithms are Maximum-in-Range and yield truthful mechanisms when paired with Vickrey-Clarke-Groves payments.

Uniform Price Auctions: Equilibria and Efficiency

We study the Uniform Price Auction, one of the standard sealed-bid multi-unit auction formats in Auction Theory, for selling multiple identical units of a single good to multi-demand bidders. Contrary to the truthful and efficient multi-unit Vickrey auction, the Uniform Price Auction encourages strategic bidding and is generally inefficient, due to a “Demand Reduction” effect; bidders tend to bid for fewer (identical) units, so as to receive them at a lower uniform price. All the same, the uniform pricing rule is popular by its appeal to the anticipation that identical items should be identically priced. Its applications include, among others, sales of U.S. Treasury notes to investors and trade exchanges over the Internet facilitated by popular online brokers. In this work, we characterize pure undominated bidding strategies and give an algorithm for computing pure Nash equilibria in such strategies. Subsequently we show that their Price of Anarchy is ee−1

Inefficiency of Standard Multi-unit Auctions

\frac {e}{e-1}

Externalities among advertisers in sponsored search

. Finally, we show that the Price of Anarchy of mixed Bayes-Nash equilibria with undominated support is at most 4−2k

Envy-Free Revenue Approximation for Asymmetric Buyers with Budgets

4-\frac {2}{k}

Tight Welfare Guarantees for Pure Nash Equilibria of the Uniform Price Auction

, where k is the number of auctioned items. To the best of our knowledge, our work provides the first (constructive) proof of existence of pure Nash equilibria in undominated strategies and the first performance evaluation (with respect to economic efficiency) of this popular auction format.

The Strong Price of Anarchy of Linear Bottleneck Congestion Games

We study two standard multi-unit auction formats for allocating multiple units of a single good to multi-demand bidders. The first one is the Discriminatory Auction, which charges every winner his winning bids. The second is the Uniform Price Auction, which determines a uniform price to be paid per unit. Variants of both formats find applications ranging from the allocation of state bonds to investors, to online sales over the internet. For these formats, we consider two bidding interfaces: (i) standard bidding, which is most prevalent in the scientific literature, and (ii) uniform bidding, which is more popular in practice. In this work, we evaluate the economic inefficiency of both multi-unit auction formats for both bidding interfaces, by means of upper and lower bounds on the Price of Anarchy for pure Nash equilibria and mixed Bayes-Nash equilibria. Our developments improve significantly upon bounds that have been obtained recently for submodular valuation functions. Also, for the first time, we consider bidders with subadditive valuation functions under these auction formats. Our results signify near-efficiency of these auctions, which provides further justification for their use in practice.

Uniform price auctions: equilibria and efficiency

We introduce a novel computational model for single-keyword auctions in sponsored search, which models explicitly externalities among advertisers, an aspect that has not been fully reflected in the existing models, and is known to affect the behavior of real advertisers. Our model takes into account both positive and negative correlations between any pair of advertisers, so that the clickthrough rate of an ad depends on the identity, relative order and distance of other ads appearing in the advertisements list. In the proposed model we present several computational results concerning the Winner Determination problem for Social Welfare maximization. These include hardness of approximation and polynomial time exact and approximation algorithms. We conclude with an evaluation of the Generalized Second Price mechanism in presence of externalities.

Discrete strategies in keyword auctions and their inefficiency for locally aware bidders

We study the computation of revenue-maximizing envy-free outcomes in a monopoly market with budgeted buyers. Departing from previous works, we focus on buyers with asymmetric combinatorial valuation functions over subsets of items. We first establish a hardness result showing that, even with two identical additive buyers, the problem is inapproximable. In an attempt to identify tractable families of the problem’s instances, we introduce the notion of budget compatible buyers, placing a restriction on the budget of each buyer in terms of his valuation function. Under this assumption, we establish approximation upper bounds for buyers with submodular valuations over preference subsets as well as for buyers with identical subadditive valuation functions. Finally, we also analyze an algorithm for arbitrary additive valuation functions, which yields a constant factor approximation for a constant number of buyers. We conclude with several intriguing open questions regarding budgeted buyers with asymmetric valuation functions.