Sarit Kraus and Jonathan Wilkenfeld. The function of time in cooperative negotiations. In Proceedings of the Ninth National Conference on Artificial Intelligence, pages 179-- 184, Anaheim, California, July 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....on strategic aspects of multi agent activity in Artificial Intelligence has grown rapidly in the recent years. Work in AI has mostly concentrated on the design of protocols for agents interaction. Hence, work in AI shares much in common with work on mechanism design in Economics (see e.g. [2, 3, 18, 4, 7, 20]) In addition, the idea of using microeconomic tools for problem solving has also been the subject of market oriented approaches to distributed planning [22] Many basic principles and ideas for the design of multi agent AI systems grew up from the mechanism design literature. Much of the ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

....49, 28] Work that follows the latter approach has been characterized by game theoretical models of negotiations. For example, Rosenschein and Genesereth [48] use such models to analyze how rational agents can coordinate their activities by striking deals among themselves. Kraus and Wilkenfeld [29] describe a situation of conflict that agents can solve by engaging into a negotiation process. The authors use classical bargaining game theory [45] where the time at which a deal is reached is an integral part of the outcome of the negotiation. Rosenschein and Zlotkin [49] advocate the use of ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

....should be executed in a situation at hand. Other work in AI include efforts on semantics of KQML [29, 41] which is closely related to earlier work of Cohen and Levesque, but it does not include the notion of value central to our approach. While work on communication in negotiation is reported in [28, 48, 49]. Recently, Tambe [42] suggested decision theoretic communication selectivity to establish mutual belief among agents in a team. This approach is similar to ours in that the focus is on whether or not an agent should transmit a given message (fact) to others. Tambe uses a decision tree containing ....

Sarit Kraus and Jonathan Wilkenfeld. The function of time in cooperative negotiations. In Proceedings of the Twelfth International Joint Conference on Artificial Intelligence, August 1991.

....negotiations as a way of reaching mutual benefit is that negotiation is a costly and time consuming process and, consequently, it may increase the overhead of coordination. In the presence of time constraints, planning and negotiation time should be taken into consideration. Kraus and Wilkenfeld [3] modeled negotiation from a cognitive standpoint. They suggest that a strategic model of negotiation takes the passage of time during the negotiation process itself in consideration. In both cases they want to prevent the agents from spending too much time on negotiation and therefore not keeping ....

S. Kraus & J. Wilkenfeld, The Function of Time in Cooperative Negotiations, In Proceedings of the 9th National Conference on Artificial Intelligence, California, pp. 179-184, 1991.

....have been used to codify and apply analyses for identifying and resolving conflicts that arise in a variety of domains, including labor negotiation[261] design integration[130] and specification integration[229] Negotiation protocol knowledge. Frameworks[39] 250] and communication protocols [133][141] 165] 191] 238] have been defined for the purpose of coordinating the sequences of messages among distributed negotiating agents. 3.2.5 Negotiation Support Systems Researchers in negotiation support systems (NSS) are addressing interactions in the context of advising a human negotiator or ....

....on criteria trade offs among decision alternatives[215] 221] 218] Using the procedure, an analyst does not explicitly 8 Many distributed artificial intelligence projects use the term negotiation to describe their work. Typically, the aspect of negotiation explored concerns explicit[38] 39] 50][133][165] 235] 237] 242] 250] or implicit[232] 233] 234] communication protocols. Such communication often involves the collaborative sharing of local aspects of a global problem. However, strategic communication in competitive environments has been explored[191] 235] Nevertheless, the generation ....

Kraus, S., Wilkenfeld, J., The function of time in cooperative negotiations, AAAI, Proceedings of Ninth National Conference on Aritficial Intelligence, July 14-19, Vol 1, pp. 179-184. Conclusions 60 1999 William N. Robinson Requirements Interaction Management GSU CIS 99-7

....The result of the negotiation process is a deal that the agents will follow. Work in AI has been mostly concerned with agents that conform to agreed upon deals. Agents may not follow irrational negotiation protocols, but will conform to deals obtained by following rational negotiation protocols (Kraus Wilkenfeld 1991; Zlotkin Rosenschein 1993; Durfee, Lee, Gmytrasiewicz 1993) 1 This differs from work in game theory (Owen 1982; Fudenberg Tirole 1991) where a joint strategy is considered unstable (and therefore unsatisfactory from a design perspective) if an agent has a rational incentive to deviate ....

....adopt only rational negotiation protocols (Zlotkin Rosenschein 1993; 2 Plans with complete information and other forms of plans will be taken as restrictions on the general form of plans considered in this paper; this point will not effect the discussion or results presented in this paper. Kraus Wilkenfeld 1991). On the other hand, work in Game Theory has been concerned with finding joint strategies which will be stable against rational deviations, where rationality is associated with expected utility maximization. We are interested in guaranteeing efficient behavior for the agents; this behavior should ....

[Article contains additional citation context not shown here]

Kraus, S., and Wilkenfeld, J. 1991. The Function of Time in Cooperative Negotiations. In Proc. of AAAI91, 179--184.

....of these protocols to the way in which information is presented. Work in mechanisms design deals with the design of multi agent interactions of rational agents. This work is fundamental to Economics [9] as well as to the attempt to design protocols for non cooperative computational environments [4, 16, 8, 5, 3, 17]. The importance of the latter has increased dramatically due to the desire to design economic interactions in the Internet setup. In such a context, the fact we have different human users in a distributed system, suggests we can not rely on standard protocol design; we have to design protocols ....

....this claim is basically correct, researchers have noticed that in order to design protocols for non cooperative computational environments, various modifications to the mechanism design theory are needed. In particular, researchers have been concerned with the fact agents are resource bounded [10, 18, 8]. However, except for the important issue of bounded resources there are several other issues that may make a difference when one tries to adapt the game theoretical approach to mechanism design in computational settings. These issues arise from the fact that the players in such a mechanism are ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

....on strategic aspects of multi agent activity in Artificial Intelligence has grown rapidly in the recent years. Work in AI has mostly concentrated on the design of protocols for agents interaction. Hence, work in AI shares much in common with work on mechanism design in Economics (see e.g. [2, 3, 18, 4, 7, 20]) In addition, the idea of using microeconomic tools for problem solving has also been the subject of marketoriented approaches to distributed planning [22] Many basic principles and ideas for the design of multi agent AI systems grew up from the mechanism design literature. Much of the research ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

....should pay a visit, then we simply need to announce our utilities for the different alternatives (assuming that we are honest about them) and this finishes our goal evaluation procedure. In the next stage, we need to decide on appropriate plans, and computational mechanisms such as negotiations [12], deals [19] or social laws ( 15] 23] may be helpful in this regard. At first glance, this solution may seem quite satisfactory, and indeed research on multi agent systems have implicitly assumed that the goal evaluation problem is solved, and concentrated on computational mechanisms for group ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

....should be executed in a situation at hand. Other work in AI include efforts on semantics of KQML [18, 29] which is closely related to earlier work of Cohen and Levesque, but it does not include the notion of value central to our approach. While work on communication in negotiation is reported in [17, 34, 35]. Communication among rational agents has also been of interest for the researchers in game theory, usually viewed a part of pre game cheap talk [6] or as threat games [20] see also discussion in [8] and references therein) These approaches, however, concentrate on the influence of pre play ....

Sarit Kraus and Jonathan Wilkenfeld. The function of time in cooperative negotiations. In Proceedings of the Twelfth International Joint Conference on Artificial Intelligence, August 1991.

....that arise in this research are: 1. When is it worth it for an agent to lie 2. How beneficial can such lies be 3. Is the system with a lying agent robust 4. How can lies be discouraged The work is an extension of the model of multi agent negotiations under time constraints presented in [7, 8, 9]. The main contributions of the current work are our ability to deal with multiple encounters among agents, and our treatment of the problem that enables elementary mutual learning. The problem of resource allocation arises when a set of agents shares a common resource (database, file, printer, ....

S. Kraus and J. Wilkenfeld. The function of time in cooperative negotiations. In Proc. of AAAI-91, pages 179--184, California, 1991.

....appeal to some central traffic controller for coordination advice, or alternatively they might engage in a negotiation resulting (say) in each robot moving slightly to its right. The use of negotiation to resolve conflicts is common in the distributed AI literature (see for example [14] 3] or [7]) Nonetheless, there are limitations to this second approach as well. By placing no constraints in advance, the number of interactions may be prohibitive; either the central arbiter may be deluged by pleas, or else agents will have to enter into negotiations at every step. While we certainly do ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

.... Introduction A number of coordination methods have been discussed in the AI literature; Some of these methods are concerned with centralized approaches to coordination [Buc89, Lan88, Stu85, TM89] Other methods are concerned with decentralized approaches [BG88] where issues such as negotiations [KW91], deals [ZR93] and consensus [DM90] play a major role. The artificial social systems methodology (e.g. BC95, MT90, ST95] attempts to bridge the gap between the fully centralized and the fully decentralized approaches to coordination. The basic idea is that the society will adopt a set of ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

....can coexist while achieving their goals. Several approaches for coordination of agent activity are discussed in the distributed systems and the DAI literature. Some examples are: protocols for reaching consensus (Dwork Moses, 1990) rational deals and negotiations (Zlotkin Rosenschein, 1993; Kraus Wilkenfeld, 1991; Rosenschein Genesereth, 1985) organizational structures (Durfee, Lesser, Corkill, 1987; Fox, 1981; Malone, 1987) and social laws (Moses Tennenholtz, 1995; Shoham Tennenholtz, 1995; Minsky, 1991; Briggs Cook, 1995) In some of these methods, the behavior of an agent is predetermined ....

....arbitrary; rational agents and reinforcement learners are the two major types of agents studied in mathematical economics, decision theory, and game theory. They are also the types of agents discussed in work in DAI which is concerned with self motivated agents (e.g. Zlotkin Rosenschein, 1993; Kraus Wilkenfeld, 1991; Yanco Stein, 1993; Sen, Sekaran, Hale, 1994) An agent s ability to function in an environment is greatly affected by its knowledge of the environment. In some special cases, we can design agents with sufficient knowledge for performing a task (Gold, 1978) but, in general, agents must ....

Kraus, S., & Wilkenfeld, J. (1991). The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pp. 179--184.

....and in distributed AI adopts a somewhat complementary view. In difference to classical work in distributed operating systems, an agent is not associated with a set of resources that it controls. The agents are autonomous entities which negotiate among themselves (Zlotkin Rosenschein, 1993; Kraus Wilkenfeld, 1991) on the use of shared resources. Alternatively, the agents (called managers in this case) may negotiate the task to be executed with the processors which may execute it (Malone, 1987) The model we adopt has the flavor of models used in distributed AI and organization theory. We assume a strict ....

Kraus, S., & Wilkenfeld, J. (1991). The function of time in cooperative negotiations. In Proc. of the 9th Nat. Conf. on Artificial Intelligence (AAAI-91), pp. 179--184.

....action of transfering some utility to the physical action of R 2 going to P2. R 1 could transfer 1 unit of utility to R 2 so that R 2 would come out even, and R 1 would still be better off. Of course, R 2 could hold out for more. This kind of negotiation has been looked at most recently by Kraus [13]. If utility is not transferable, or if R 1 is more greedy (and willing to take some risk) then R 1 could instead threaten R 2 with not going to P1 unless R 2 goes to P2. This could be an effective strategy, but only if R 2 takes the threat seriously. Issues of how threats can be made more ....

Sarit Kraus and Jonathan Wilkenfeld. The function of time in cooperative negotiations. In Proceedings of the Twelfth International Joint Conference on Artificial Intelligence, August 1991.

....Moreover, in some cases the cost of avoiding conflicts of a particular type may be higher than the cost of resolving them once they occur. A comprehensive social system must therefore also contain a component that describes how conflicts are to be handled once they occur. The work on negotiations [18], deals [30] consensus [27] interaction protocols (e.g. 2] as well as many other forms of conflict resolution can be viewed as handling this very delicate and complex aspect of the design of a social system. 4 Logical Reasoning about Social Systems The previous sections introduced and ....

....to play in this context, so long as sufficient care is taken in defining, studying, and applying social systems to multi agent activity. Much work has been devoted to an explicit and formal study of the centralized approach and of the on line resolution of conflicts (e.g. 19] 15] 37] 30] [18]) Our work is the first to discuss explicitly and formally the computational mechanism that applies for noncentralized intermediate solutions. In spite of the generality of our work, we wish to emphasize that our work in no way diminishs the crucial importance of mechanisms for interaction and ....

S. Kraus and J. Wilkenfeld. The Function of Time in Cooperative Negotiations. In Proc. of AAAI-91, pages 179--184, 1991.

....in their paper Cooperation and Conflict Resolution via Negotiation Among Autonomous Agents in Non Cooperative Domains, also employ formal methods for investigating coordination among agents. Their approach takes a decision theoretic, probablistic view of decisionmaking in multiagent environments [24, 36, 50], where each agent is acting so as to maximize its expected utility. By considering the worths of their goals and the costs of achieving them, the interaction between agents can be characterized as requiring cooperation, compromise, or conflict. Moreover, by utilizing probablistic methods, the ....

Sarit Kraus and Jonathan Wilkenfeld. The function of time in cooperative negotiations. In Proceedings of the Twelfth International Joint Conference on Artificial Intelligence, August 1991.

....E40 179 Cambridge, MA 02139 Gilad mit.edu Note: This article appears in Artificial Intelligence journal 75(2) 297 345, 1995. This material is based upon work supported by the National Science Foundation under Grant No. IRI9123460. Some material in this paper appeared in preliminary form in [26, 27]. We thank Karen Lochbaum for her comments. y Also affiliated with the Institute for Advanced Computer Studies, University of Maryland, College Park. z This research was done while the author was at the Computer Science Department in the Hebrew University and was supported by Leibniz Center for ....

S. Kraus and J. Wilkenfeld. The function of time in cooperative negotiations. In Proc. of AAAI-91, pages 179--184, California, 1991.

....function, and that rational behavior involves maximizing its expected utility. Our work belongs to the DAI class of Multi Agent Systems(MA) e.g. 52, 61, 63, 65] rather than to the Distributed Problem Solving (DPS) class (e.g. 58, 17, 41] as discussed in Section 2 below. In previous work [31, 32, 34] we have developed a formal strategic model of negotiation that takes into consideration the passage of time during the negotiation process. In this paper, we extend this model to deal with incomplete information in multiple encounters when the agents negotiate on sharing a resource. There may be ....

....paper we assume that there is one agent that is currently using a resource (A symbolizing access ) and that there is another agent which also wants to use it (W symbolizing waiting ) W wishes to gain access to the resource during the next M time periods. First we modify several assumptions of [31, 32, 34] to fit a situation in which agents may be of several types. For i 2 Type, if the type of agent A (respectively W ) is i, we denote it by A i (respectively W i ) If a condition holds regardless of the type of agent A (respectively W ) we use A (respectively W ) The first assumption states that ....

S. Kraus and J. Wilkenfeld. The function of time in cooperative negotiations. In Proc. of AAAI-91, pages 179--184, California, 1991.

....that communicate and cooperate. For example, Sycara [105, 104] presents a model of negotiation that combines case based reasoning and optimization of the multi attribute utilities. This model is used 10 in labor management negotiations where two agents need to reach an acceptable agreement. In [50, 51, 52], a strategic negotiation model is presented for situations where a set of selfmotivated autonomous agents have common goals that they want to satisfy as soon as possible. Each agent, while wanting to minimize its costs, prefers to do as little as possible and therefore tries to reach an agreement ....

.... will decide how much effort to expend, but its decision may be influenced by the contract offered 7 In our previous work on negotiation under time constraints, we have identified perfect equilibrium strategies and proposed to develop a library of meta strategies to be used when appropriate [50, 51, 52]. 15 Notation Meaning Comments Effort Set of efforts of the contractor e; e 1 ; e i 2 Effort Outcome Set of possible monetary outcomes q; q 1 ; q j 2 Outcome. q(e) 2 Outcome for carrying out a task. when q is a function of e 2 Effort Rewards Set of possible monetary rewards to r; r 1 ....

S. Kraus and J. Wilkenfeld. The function of time in cooperative negotiations. In Proc. of AAAI-91, pages 179--184, California, 1991.

....Studies, University of Maryland, College Park, MD 20742, USA Osher Yadgar Dept. of Math and CS, Bar Ilan University, Ramat Gan, 52900 Israel Abstract Cooperation among autonomous agents has been discussed in the DAI community for several years. Papers about cooperation [6,45] negotiation [33], distributed planning [5] and coalition formation [28,48] have provided a variety of approaches and several algorithms and solutions to situations wherein cooperation is possible. However, the case of cooperation in large scale multi agent systems (MAS) has not been thoroughly examined. ....

S. Kraus and J. Wilkenfeld. The function of time in cooperative negotiations. In Proc. of AAAI-91, pages 179--184, California, 1991.

....its own sub tasks (for example, they may need to share the same resources) but their overall task is the same. Our work takes as a point of departure the work of researchers who have studied the negotiations that could take place among agents that serve the interests of truly distinct parties [59, 47, 63, 25, 30, 31]. The agents are autonomous; they have their own utility functions, and no global notion of utility plays a role in their design. The agents are individually motivated. For example, Sycara [59] presented a model of negotiation that combines case based reasoning and optimization of the ....

....satisfying A0 A5 such that jc P j c I 1 and P is the first player. If it is I s turn and s P;T 1 I 0 then g P;T I s P;T . 7 This is the situation when autonomous agents negotiate over the sharing of a common resource, and one of them is using the resource during the negotiation ([30]) In the next lemma we show that if I offers P something less preferable by P than s P;t , P will opt out. Lemma 11 Let ( f ; g) be a P.E. of a model satisfying A0 A5 such that jc P j c I 1 and P is the first player. If T is odd (i.e. I s turn) and s P;T 1 I 0 then if g T I ....

S. Kraus and J. Wilkenfeld. The function of time in cooperative negotiations. In Proc. of AAAI-91, pages 179--184, California, 1991.

....to expend, but its decision may be influenced by the contract offered by the contracting agent. 2 In our previous work on negotiation under time constraints, we have identified perfect equilibrium strategies and proposed to develop a library of meta strategies to be used when appropriate [ Kraus and Wilkenfeld, 1991a; Kraus and Wilkenfeld, 1991b; Kraus et al. 1994 ] Notation Meaning Comments Effort Set of efforts of the contracted agent e; e 1 ; e i 2 Effort Outcome Set of possible monetary outcomes q; q 1 ; q j 2 Outcome. q(e) 2 Outcome of carrying out a task. when q is a function of e 2 ....

....may be influenced by the contract offered by the contracting agent. 2 In our previous work on negotiation under time constraints, we have identified perfect equilibrium strategies and proposed to develop a library of meta strategies to be used when appropriate [ Kraus and Wilkenfeld, 1991a; Kraus and Wilkenfeld, 1991b; Kraus et al. 1994 ] Notation Meaning Comments Effort Set of efforts of the contracted agent e; e 1 ; e i 2 Effort Outcome Set of possible monetary outcomes q; q 1 ; q j 2 Outcome. q(e) 2 Outcome of carrying out a task. when q is a function of e 2 Effort Rewards Set of ....

S. Kraus and J. Wilkenfeld. The function of time in cooperative negotiations. In Proc. of AAAI-91, pages 179--184, California, 1991.

No context found.

Sarit Kraus and Jonathan Wilkenfeld. The function of time in cooperative negotiations. In Proceedings of the Ninth National Conference on Artificial Intelligence, pages 179-- 184, Anaheim, California, July 1991.

First 50 documents

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