Eithan Ephrati and Jeffrey S. Rosenschein. Deriving consensus in multi-agent systems. Artificial Intelligence. To appear.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....Lesser [17] to selfinterested agents. In these systems, agents communicate and negotiate directly with each other. In MAGNET, on the other hand, agents interact with each other through an independent market infrastructure. Mechanisms to reduce counterspeculation, such as the Clarke tax mechanism [6] or the Vickrey auction have been proposed for automated negotiation of self interested agents. In Sandholm s TRACONET system [18] both the bidding and contract execution mechanisms are complicated by the need to operate in an environment where agents cannot trust each other. He does not assume or ....

E. Ephrati and J. Rosenschein. Deriving consensus in multiagent systems. Artificial Intelligence, 87:21--74, 1996.

....assumed to be created by the same designer (or group of designers) and thus, work together to solve common goals. In MASs, however, agents may have their own private goals and act selfishly towards the achievement of these goals. Therefore each agent may have its private profile of preferences [ 10 ] Three kinds of entities, including tasks, information and resources, are necessary in DAI. Tasks are the problems to be solved. They can be roughly divided into two types: i) questions to be answered and (ii) actions to be performed. Information is data (either static or dynamic) to describe ....

.... several notable efforts for describing autonomy, e.g. artificial social systems [ 24 ] social laws [ 33 ] establish cooperation through the formalization of agents intentions [ 12 ] 13 ] reaching consensus by negotiation [ 4 ] 8 ] 17 ] 28 ] 34 ] or by economic decision process [ 10 ] and building ideal rational agents [ 29 ] For this kind of cooperation, the information level cooperation is important, because the information an agent receives or sends is just based on its belief and interests. At this stage, when an agent gathers this information, it first should fuse ....

E. Ephrati and J. S. Rosenschein, Deriving consensus in multiagent systems, Artificial Intelligence, 1996, 87(1-2): 21-74.

....# , and # # is the weight of the #th dimension. One major weakness of this method is its susceptibility to small variations in the stated preferences. Voting systems have previously been applied to multiagent system domains [Ephrati and Rosenschein, 1991; Ephrati et al. 1994; Rosenschein, 1995; Ephrati and Rosenschein, 1996] but not in the context of cooperative voters. Ephrati et al. s approach focuses on voting by all users of a meeting to reach consensus on an acceptable time for the meeting and is a non cooperative voting scenario where voters can vote insincerely [Ephrati et al. 1994 ] They can therefore ....

Eithan Ephrati and Jeffrey S. Rosenschein. Deriving consensus in multiagent systems. Artificial Intelligence, 87(1-2):21--74, November 1996.

....of these rules is to allow the agents to make constructive agreements. Their analysis assumes that the negotiating agents have similar capabilities. The protocol we present in this paper does not require that assumption. Mechanisms to reduce counterspeculation, such as the Clarke tax mechanism (Ephrati Rosenschein 1996) or the Vickrey auction (Vickrey 1961) have been proposed for automated negotiation of self interested agents. The architecture we present can support the Vickrey auction, and eliminates one of its limitations by providing a structure that can act as a trusted auctioneer (Sandholm 1996) A ....

Ephrati, E., and Rosenschein, J. 1996. Deriving consensus in multiagent systems. Artificial Intelligence 87:21--74.

....when they actually have an influence. In our case, each data item is unique, so a competitive approach cannot be used. Furthermore, in our model, there are few servers involved, so the competitive assumption does not hold. Thus, we will use negotiations for our problem. Ephrati and Rosenschein [17] suggest using the Clarke Tax voting procedure in order to reach a consensus in MA systems. In this procedure, each agent expresses its cardinal utility for each possibility, and the one with the highest sum of utilities is selected. Taxes are taken from the agents in order to ensure truthful ....

....does not know the mean usage of each dataset by clients of each area. Each server knows only the past usage of the datasets stored locally, and it also knows the past usage of datasets by clients in its area. There are several approaches to dealing with incomplete information decision making [32, 17, 41], but each of them has several limitations when applied to complex problems 8 such as the data allocation problem. In [51] we describe some approaches and discuss their limitations for our domain. However, the strongest limitations are the problem of applying the protocols of these approaches to ....

E. Ephrati and J. S Rosenschein. Deriving consensus in multiagent systems. Artificial Intelligence, 87(1--2):21--74, 1996.

....of these rules is to allow the agents to make constructive agreements. Their analysis assumes that the negotiating agents have similar capabilities. The protocol we present in this paper does not require that assumption. Mechanisms to reduce counterspeculation, such as the Clarke tax mechanism [2] or the Vickrey auction [19] have been proposed for automated negotiation of self interested agents. The architec A Market Architecture for Multi Agent Contracting April 18, 1997 2 ture we present can support the Vickrey auction, and eliminates one of its limitations by providing a structure ....

E. Ephrati and J. Rosenschein, Deriving consensus in multiagent systems, Artificial Intelligence, Vol 87, 1996, 21-74.

....mechanism exists. This mechanism is the famous Clarke s mechanism [3] Work on multi agent systems in AI [1, 4, 12, 2] share with work in information economics similar concerns, and therefore there is no doubt that the introduction of the Clarke s mechanism is of fundamental importance to AI too [5]. One major issue that work in AI and Computer Science need to address when applying the Clarke s mechanism is concerned with the fact that the participants (the agents and the center) are resource bounded. Discussions on resource bounds in the context of mechanism design are rare. Exceptions ....

....concerned with the fact that the participants (the agents and the center) are resource bounded. Discussions on resource bounds in the context of mechanism design are rare. Exceptions include work on resource bounds in the context of auctions [13] and in the context of coalition formation [15] In [5], Ephrati and Rosenschein deal with the communication burden in the context of the Clarke s mechanism. However, as we show, there is a major obstacle in applying the Clarke s mechanism even in very simple settings. Formally, this obstacle is captured by the fact that the problem of optimizing ....

E. Ephrati and J. Rosenschein. Deriving consensus in multi-agent systems. Artificial Intelligence, 87, 1996.

....This rule is designed to strengthen the Borda count voting rule: if there is a Condercet Winner, then choose it, else apply the Borda count voting rule. Voting systems have previously been applied to distributed artificial intelligence domains [Ephrati and Rosenschein, 1991, Rosenschein, 1995, Ephrati and Rosenschein, 1996] and meeting schedulers in particular [Ephrati et al. 1994] but not in the context of cooperative voters. Ephrati et al. s approach focuses on voting by all users of a meeting to reach consensus on an acceptable time for the meeting and is a non cooperative voting scenario where voters can ....

Eithan Ephrati and Jeffrey S. Rosenschein. Deriving consensus in multiagent systems. Artificial Intelligence, 87(1-2):21--74, November 1996.

....of these rules is to allow the agents to make constructive agreements. Their analysis assumes that the negotiating agents have similar capabilities. The protocol we present in this paper does not require that assumption. Mechanisms to reduce counterspeculation, such as the Clarke tax mechanism [2] or the Vickrey auction [17] have been proposed for automated negotiation of self interested agents. The architecture we present can support the Vickrey auction, and eliminates one of its limitations by providing a structure that can act as a trusted auctioneer [12] A variety of architectures ....

E. Ephrati and J. Rosenschein. Deriving consensus in multiagent systems. Artificial Intelligence, 87:21--74, 1996.

No context found.

Eithan Ephrati and Jeffrey S. Rosenschein. Deriving consensus in multi-agent systems. Artificial Intelligence. To appear.

No context found.

: 579-594, 1982. #8# Ephrati, E., and Rosenschein, J., Deriving Consensus in Multiagent Systems, Arti#cial Intelligence

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