Esther David

Self-organized routing for wireless microsensor networks

In this paper, we develop an energy-aware self-organized routing algorithm for the networking of simple battery-powered wireless microsensors (as found, for example, in security or environmental monitoring applications). In these networks, the battery life of individual sensors is typically limited by the power required to transmit their data to a receiver or sink. Thus, effective network-routing algorithms allow us to reduce this power and extend both the lifetime and the coverage of the sensor network as a whole. However, implementing such routing algorithms with a centralized controller is undesirable due to the physical distribution of the sensors, their limited localization ability, and the dynamic nature of such networks (given that sensors may fail, move, or be added at any time and the communication links between sensors are subject to noise and interference). Against this background, we present a distributed mechanism that enables individual sensors to follow locally selfish strategies, which, in turn, result in the self-organization of a routing network with desirable global properties. We show that our mechanism performs close to the optimal solution (as computed by a centralized optimizer), it deals adaptively with changing sensor numbers and topology, and it extends the useful life of the network by a factor of three over the traditional approach.

Bidding in sealed-bid and English multi-attribute auctions

In this paper we consider an extension of the traditional auction mechanism, the multi-attribute auction, which enables negotiation on several attributes in addition to the price of the item. In particular, we consider a procurement auction in which the buyer is the auctioneer and the sellers are the bidders. Such domains include auctions on task allocation, services, etc. We focus on three auction protocols for the case of multi-attribute items; a variation of the first-price sealed-bid protocol termed first-score sealed-bid, a variation of the second-price sealed-bid protocol termed second-score sealed-bid, and a variation of the English auction protocol termed sequential full information revelation. We analyze a specific model for these protocols and we provide optimal and stable strategies for the auctioneer agent and for the bidder agents participating in multi-attribute auctions. In addition, we analyze the auctioneers/buyers expected payoff and suggest an optimal scoring rule to be announced according to the protocol. Finally, we reveal that the buyers expected payoff in all three protocols, the first-score-sealed-bid auction, the second-score sealed-bid auction and the English auction, differ only by a predefined constant. We prove that the optimal scoring rule is equal in all three protocols. This result can be interpreted as the extension of the equivalence theory of the single attribute for the case of multi-attribute items.

An English Auction Protocol for Multi-attribute Items

In this paper, we suggest using an English Auction Protocol for a procurement multi-attribute auction in which the item for sale is defined by several attributes, the buyer agent is the auctioneer, and the seller agents are the bidders. Such domains include auctions on task allocation, services, or compound products. At the beginning of the auction the buyer agent announces the required properties of the item, and then various seller agents propose bids, which are composed of specific configurations that match its request. Each proposed bid should be better for the buyer agent than the previous bid, w.r.t. the announced requirements of the buyer agent. Finally, the last suggested bid will win, and the seller agent that suggested this bid will be committed to it. We consider two utility function models for the English auction protocols and provide the optimal bidding strategies for the seller agents and the optimal auction design for the buyer agents regarding both models.

Protocols and strategies for automated multi-attribute auctions

In this paper, we consider a model of a procurement multi-attribute auction in which the sales item is defined by several attributes, the buyer is the auctioneer, and the sellers are the bidders. Such domains include auctions on task allocation, services, or compound items. The buyer announces a scoring rule, according to its preferences, before the auction starts, and each seller places a bid, which describes the attributes of the item it offers for sale.First, we consider a variation of the first-price sealed-bid protocol, and we provide optimal and stable strategies for the buyer agent and for the seller agents participating in the multi-attribute auction. In addition, we analyze the buyers expected revenue and suggest an optimal scoring rule that can be announced. Second, we consider four variations of the English auction for the case of a multi-attribute item, and we prove that, given some assumptions, they all converge to the same result. We also discuss which variation is preferred for different types of environments. Moreover, we show under which conditions, announcing the truth about buyer preferences is the optimal strategy for the buyer.

An advanced bidding agent for advertisement selection on public displays

In this paper we present an advanced bidding agent that participates in first-price sealed bid auctions to allocate advertising space on BluScreen - an experimental public advertisement system that detects users through the presence of their Bluetooth enabled devices. Our bidding agent is able to build probabilistic models of both the behaviour of users who view the adverts, and the auctions that it participates within. It then uses these models to maximise the exposure that its adverts receive. We evaluate the effectiveness of this bidding agent through simulation against a range of alternative selection mechanisms including a simple bidding strategy, random allocation, and a centralised optimal allocation with perfect foresight. Our bidding agent significantly outperforms both the simple bidding strategy and the random allocation, and in a mixed population of agents it is able to expose its adverts to 25% more users than the simple bidding strategy. Moreover, its performance is within 7.5% of that of the centralised optimal allocation despite the highly uncertain environment in which it must operate.

Evaluating BluScreen: Usability for Intelligent Pervasive Displays

Within a ubiquitous environment, market-based approaches can be used to select the most appropriate material for a public display, depending on factors such as the audiences preferences and diversity of interest. Likewise, strategies used by agents to compete for customer attention should strive to be rational, based on contextual observations of user-preferences within the local environment and include a reward mechanism based on audience responses. But while such systems currently exist, utilizing Bluetooth-enabled mobile phones to uniquely identify and detect the presence of individuals within a localised environment, there is little known about their effectiveness, or even how to assess usability for these systems. In this paper, we present the details a user study that contributed to the development of an interaction model that supports a structured methodology for evaluating intelligent pervasive displays.

Learning environmental parameters for the design of optimal english auctions with discrete bid levels

In this paper we consider the optimal design of English auctions with discrete bid levels. Such auctions are widely used in online internet settings and our aim is to automate their configuration in order that they generate the maximum revenue for the auctioneer. Specifically, we address the problem of estimating the values of the parameters necessary to perform this optimal auction design by observing the bidding in previous auctions. To this end, we derive a general expression that relates the expected revenue of the auction when discrete bid levels are implemented, but the number of participating bidders is unknown. We then use this result to show that the characteristics of these optimal bid levels are highly dependent on the expected number of bidders and on their valuation distribution. Finally, we derive and demonstrate an online algorithm based on Bayesian machine learning, that allows these unknown parameters to be estimated through observations of the closing price of previous auctions. We show experimentally that this algorithm converges rapidly toward the true parameter values and, in comparison with an auction using the more commonly implemented fixed bid increment, results in an increase in auction revenue.

Selfish Sensors in Wireless Micro-Sensor Networks

In this paper we develop an energy aware decentralised routing algorithm for adhoc networking of battery-powered wireless microsensors. The useful life of such networks is limited by the battery life of individual sensors and thus the goal of any networking algorithm is to maximise both the lifetime and the coverage of the network, whilst dealing adaptively with sensor failures and changes in network topology. As sensors may be owned and supported by different stakeholders, we view them as selfish agents maximising their own utility. To this end, we develop a mechanism that enables such agents to follow locally selfish strategies which, in turn, result in the achievement of good global performance.

Agents for Information Broadcasting

In this paper we consider an environment which consists of one broadcasting entity (producer) which broadcasts information to a large number of personal computer users, who can down-load information to their PC disks (consumers). We concentrate on the most critical phase of the broadcasting system operation, which is the characterization of the users’ needs in order to maximize the efficiency of the broadcast information. Since the broadcasting system can not consider each user in isolation, it has to consider certain communities of users. We have proposed using a hierarchic distributed model of software agents to facilitate receiving feedback from the users by the broadcasting system. These agents cluster the system’s users into communities with similar interest domains. Subsequently, these agents calculate a representative profile for each community. Finally, the broadcasting agent builds an appropriate broadcasting program for each community. We developed a simulation of the broadcasting environment in order to evaluate and analyze the performance of our proposed model and techniques. The simulation results support our hypothesis that our techniques provide broadcasting programs, which are of great interest to the users.

No More Damaging Ads on Your Own Webpage -- Strategy Proof Mechanisms for Ad Placement

Recently we have been witnessing the phenomena of private Web site owners who are willing to dedicate space on their site for advertisements managed by leading Web companies (e.g., Google, Yahoo). In most cases the choice of the advertisements displayed on a certain site is made without taking into account the preferences of the Web site. In this paper, we focus on the design of mechanisms that are beneficial for both the Web site owner, by allowing placement of highly desired ads, and the advertisers, by introducing a dispensation scheme to reduce the prices paid for highly relevant ads. In this paper, we analyze the case of multi-slots offered by a certain Web site, and we offer three strategy-proof mechanisms that differ in their ranking function, allocation rules, and payment schemes. The performance of all the mechanisms are compared and discussed.