Noam Nisan and I Segal. The communication complexity of efficient allocation problems. Technical report, Hebrew University and Stanford University, 2002.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... incrementally elicits the bidders preferences by querying them (based on the preference information elicited so far) 6, 8, 7] In the worst case, an exponential amount of communication is re quired to allocate the items even approximately optimally, if the bidders can have general preferences [10]. This holds even when bidders can dispose of extra items for free, that is, their valuation functions are monotone. However, experimentally, preference elicitation appears to help quite a bit [9] Nonetheless, the amount of querying can be prohibitively large when the bidders have general ....

N. Nisan and I. Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th.

.... elicitation for non price based combinatorial auctions [HS02, CS01] Also related, via oracle based models of valuation queries, is a recent result that proves that the worst case communication complexity is exponential in all efficient combinatorial auctions, whether iterative or sealed bid [NS02]. Proxy bidding agents were previously proposed in the context of combinatorial auctions [PU00b] and were recently revisited in the context of combinatorial auctions by Ausubel Milgrom [AM02] although there the motivation was primarily to simplify strategic analysis; one shot revelation to ....

Noam Nisan and I Segal. The communication complexity of efficient allocation problems. Technical report, Hebrew University and Stanford University, 2002.

....another problem, because the number of bundles is exponential in the number of items and an agent can have quite general preferences across bundles. The worst case communication complexity of a fully efficient combin tori uction is exponential, both for direct nd indirect revelation mechanisms [24], lthough indirect mechisms cn chieve better vergecse performce. One pproch to ddress the communication complexity is to develop structured bidding lnguges that re compact for pticul gent preferences [22, 6] An other pproch is to place explicit restrictions on the expressivity of lguge, while ....

N. Nisan and I. Segal. The communication complexity of efficient allocation problems. Technical report, Hebrew University and Stanford University, 2002.

....for example in the context of ascending combinatorial auctions [33, 4] Of course, the transformation can also change other properties, such as a mechanism s communication complexity. The communication complexity properties of mechanisms for combinatorial auctions has been studied elsewhere [16, 30], and will not be the focus of this paper. We already gave an example in the Introduction of the effect that this direct revelation transformation can have on preference elicitation costs. While an agent can implement the equilibrium strategy in the staged second price auction from incomplete ....

....Z [ is a mapping from the possible types of agent to a subset of types. Given a query, then the response, An oracle based communication model can be interpreted as a model of costly preference elicitation, with a suitably expressive oracle. However, the current work of Nisan Segal [30] on communication complexity corresponds to a so called black box model, which allows an agent to ask an arbitrary question about preferences. Y from an agent, makes the claim that the agent s type is within the set K , e.g. my value bounds are [5,10] indicates that ....

Nisan, N. and I. Segal: 2002, `The communication complexity of efficient allocation problems'. Technical report, Hebrew University and Stanford University.

.... avenue involves studying more sophisticated equilibrium notions which take into account that players have limited memory (e.g. 1, 14, 32, 39, 41] or limited capability to solve optimization problems (e.g. 19, 23, 24, 34] There are also open issues on communication complexity in games (e.g. [7, 8, 17, 35, 50]) and on the complexity of computing general equilibria ( market equilibria ) e.g. 50] and other solutions. There are numerous open research questions even in the area of computing solutions to noncooperative games. Some recent work has focused on novel knowledge representations which, in ....

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th. 12

....revealed it is, however, a challenge to reduce this amount whenever possible. The algorithm was designed with this objective. In general, preference communication in combinatorial auctions is provably exponential (even to find an approximately optimal solution) in the theoretical worst case [16]. This may be especially useful in settings where the communication between the bidder and the auctioneer is public. We will also discuss the computation of Vickrey payments based on the information collected while executing the algorithm. We show that differential information suffices to ....

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th.

....We make the usual assumption that the valuations are measured in money (dollars) and thus can be directly compared. A recent theoretical result shows that even with free disposal, in the worst case, finding an (even only approximately) optimal allocation requires exponential communication [12]. 3 Inference and constraint network The elicitor, as we designed it, never asks a query whose answer could be inferred from the answers to previous queries. To support the storing and propagation of information received from the agents, we have the elicitor store its information in a constraint ....

N. Nisan and I. Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th.

....in the allocation. The goal of the elicitor is to determine an optimal allocation with as little elicitation as possible. A recent theoretical result shows that even with free disposal, in the worst case, finding an (even only approximately) optimal allocation requires exponential communication [14]. Therefore, we will judge the techniques successful if they reduce communication from full revelation by an asymptotic amount. 3 Elicitor s inference and constraint network The elicitor, as we designed it, never asks a query whose answer could be inferred from the answers to previous queries. ....

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th.

.... payments and that all information necessary to compute the payments is already available once the algorithm has determined an ef In general, preference communication in combinatorial auctions is provably exponential (even to find an approximately optimal solution) in the theoretical worst case [14]. This may be especially useful in settings where the communication between bidder and auctioneer is public. ficient allocation. Consider the following example with unit demand consumers and a benevolent auctioneer. 2 Example of Differential Elicitation Rank 3 2 1 4 Agent 2 2 5 3 0 Rank ....

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th.

.... elicits the bidders preferences by querying them (based on the preference information elicited so far) 6, 8, 7] In the worst case, an exponential amount of communication is required to even approximately optimally allocate the items to the bidders if the bidders can have general preferences [10]. This holds even when bidders can dispose of extra items for free, that is, their valuation functions are monotone. However, in practice only a vanishing fraction of the bidders preferences needs to be elicited before the optimal allocation of items to bidders can be determined [9] Despite ....

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th.

....in the allocation. The goal of the elicitor is to determine an optimal allocation with as little elicitation as possible. A recent theoretical result shows that even with free disposal, in the worst case, finding an (even only approximately) optimal allocation requires exponential communication [14]. Therefore, we will judge the techniques successful if they reduce communication from full revelation by an asymptotic amount. 3 Elicitor s inference and constraint network The elicitor, as we designed it, never asks a query whose answer could be inferred from the answers to previous queries. ....

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems, 2002. Draft. Second version March 5th.

No context found.

Noam Nisan and I Segal. The communication complexity of efficient allocation problems. Technical report, Hebrew University and Stanford University, 2002.

No context found.

N. Nisan and I. Segal. The communication complexity of efficient allocation problems. Technical report, Hebrew University and Stanford University, 2002.

No context found.

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems Draft. March, 2002.

No context found.

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems. Draft. Second version March 5th, 2002.

No context found.

Nisan, N., Segal, I.: The communication complexity of efficient allocation problems (2002) Draft. Second version March 5th.

No context found.

Noam Nisan and Ilya Segal. The communication complexity of efficient allocation problems. Draft. Second version March 5th, 2002.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC