Jennings, N. R.; and E. H. 1992. "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments." In Proceedings of the Tenth National Conference on Artificial San Jose, California, 269-275.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....failure rates, as indicated by extensive performance measurements. 2.0 COOPERATING DIAGNOSIS SYSTEMS IN A DISTRIBUTED ARCHITECTURE The need for robust mechanisms of cooperation among real time diagnostic modules has been an important driver of the system architecture. The notion of joint y [Jennings and 1992] as an alternative to the more conventional notion of agents acting in self interest 1988, Cohen 1990] has been amended with modular problem decomposition and data driven reasoning in order to minimize the need for communication between agents. The various modules in the distributed architecture ....

Jennings, N. R.; and E. H. 1992. "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments." In Proceedings of the Tenth National Conference on Artificial San Jose, California, 269-275.

....to achieve an order of magnitude increase in performance. 2.0 COOPERATING EXPERT SYSTEMS EMBEDDED IN A DISTRIBUTED ARCHITECTURE Recently, the need for mechanisms of cooperation that are sufficiently robust for real world monitoring applications has become a research driver. Systems such as GRATE [Jennings and 1992] contribute toward a clearer and more easily implementable interaction of agents during collaborative problem solving. GRATE addresses a problem domain in which events occur unpredictably and decisions may be based on incomplete or imprecise data. Toward this end, the notion of joint ....

Jennings, N. R.; and E. H. 1992. "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments." In Proceedings the Tenth National Conference on Artificial Intelligence, San Jose, California. 269-275.

....responsibility, commitment, authority and independence) with respect to each goal held by a respective agent. 4. RELATED WORK Several researchers have addressed the development of adaptive, intelligent, agent based systems. These efforts range from representing coordination and collaboration [1], to production rule based systems for organizational self design [2] to partial global planning [3] to the development of multiagent planning architectures [4] and many others. The majority of previous work has focused on providing architectures with a constrained set of control schemes for ....

....of formal studies of large scale dynamic systems [14] and investigations of dynamic creation and destruction of agents [16] The mechanics of changing organizational structure is also an important area of study . Several researchers have investigated how agents can form problem solving groups [17][1]. In addition, centralized, distributed, and group approaches have been compared [18] 19] These studies provide important insight for autonomy reasoning. Gasser and Ishida [2] also provide an understanding of both adaptive self configuration (through Organization SelfDesign) and tradeoffs ....

Jennings, N. R., and Mamdani, E. H. Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments. In Proceedings of the Tenth National Conference on Artificial Intelligence, pp. 269-275. American Association for Artificial Intelligence 1992.

....goals typically handled by receptionists at the front desks of buildings on the Microsoft corporate campus. Receptionists serve a vital role in facilitating daily activities at each building at Microsoft. When people interact with the receptionist, they engage in a joint activity (Clark, 1996; Jennings and Mamdani, 1992; Levinson, 1992) A joint activity is a task oriented, social event with constraints on participants, setting, and most of all, reasonable or allowable contributions. Participants in a joint activity assume that they share some common set of beliefs about the activity, including assumed roles and ....

Jennings, N.R., and Mamdani, E.H. (1992). Using joint responsibility to coordinate collaborative problem solving in dynamic environments. Proceedings of the Tenth National Conference on Artificial Intelligence. Menlo Park: AAAI Press. 269-275.

.... Sandell 1981a] and investigations of dynamic creation and destruction of agents [Ishida, Gasser, Yokoo 1992] The mechanics of changing organizational structure is also an important area of study. Several researchers have investigated how agents can form problem solving groups [Numaoka 1992, Jennings Mamdani 1992]. Gasser and Ishida [1991] provide an understanding of both adaptive self configuration (through Organization Self Design) and tradeoffs corresponding to different organizational structures. Approach Figure 3 depicts an agent architecture, which supports dynamic adaptive autonomy. Each agent ....

....responsibility distribution for a goal completely determines the autonomy level assignment to that goal. Therefore, in order to change a goal s level of autonomy, the agent must change its responsibility distribution. Responsibility has been previously recognized as a way to guide agent actions [Jennings 1992] and can be computationally linked to agent identity, states, and methods [Schlenker et al. 1994] The other three autonomy level constructs supply additional meaning which is necessary to reason about preferred responsibility distributions and their corresponding autonomy levels. Responsibility ....

Jennings, N. R., and Mamdani, E. H. 1992. Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments. In Proceedings of the Tenth National Conference on Artificial Intelligence, 269-275. American Association for Artificial Intelligence.

....An important characteristic of industrial control applications is the level of dynamicity they exhibit. Due to this non deterministic nature it is difficult to specify how the system should behave in all cases in advance. It is desirable that the Agents apply general techniques at run time [4]. We present sophisticated cooperation protocols for negotiation such as contract net where an Agent broadcasts a request to a group of acquaintances, the recipients then respond with a bid indicating the earliest time at which they can perform it and when all bids have been received the ....

- N. Jennings & A. Mamdani "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments" AAAI, San Jose,

....exchanging private views of tasks, communicating results, handling various types coordination relationships) enable agents to modify their subjective view of the task structure and their commitments to tasks in the task structure, ultimately improving performance. The Joint Responsibility model [23] prescribes when and how agents should form teams and how team members should behave during joint action. The code of conduct imposed by Joint Responibility ensures that the group will operate in a coordinated and efficient manner and that it is robust in face of changing circumstances. Given the ....

N. R. Jennings and E. Mamdani. Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments. In Proceedings of 10-th National Conference on AI, San Jose, CA, pp 269-275, 1992.

....exchanging private views of tasks, communicating results, handling various types coordination relationships) enable agents to modify their subjective view of the task structure and their commitments to tasks in the task structure, ultimately improving performance. The Joint Responsibility model [32] prescribes when and how agents should form teams and how team members should behave during joint action. The code of conduct imposed by Joint Responibility ensures that the group will operate in a coordinated and efficient manner and that it is robust in face of changing circumstances. Given the ....

N. R. Jennings and E. Mamdani. Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments. In Proceedings of 10-th National Conference on AI, San Jose, CA, pp 269-275, 1992.

.... for business processes [12] and power grid monitoring [17] The problem of modelling the activity of teams of artificial agents [14] 20] is a combination of two sub problems: the first is the modelling of the team itself [10] 38] 43] and the second is the modelling of the team activity [18], 20] The problem of modelling teams and team behaviour is very complex and includes a wide range of sub problems, in particular representing teams with a variety of organizational structures and providing a representation of team behaviour that 2 allows the implementation of a variety of ....

....and decision making processes are implicit in the team tactics and that each sub team is responsible to inform other sub teams of any change to the tactics. This type of behaviour corresponds to the notions of commitment cf. Cohen and Levesque [5] and responsibility cf. Jennings and Mamdani [18]. 4.3. Implementation In dMARS the agent s beliefs are implemented as relations (or predicates) in a relational database. The organizational structures that a team has adopted are represented as a relation between the team and the organizational structure. We refer to this relation as ....

N. R. Jennings and E. H. Mamdani. Using joint responsibility to coordinate collaborative problem solving in dynamic environments. In Proceedings of the Tenth National Conference on Artificial Intelligence, pages 269--275, San Jose, CA, USA, 1992.

....may profitably be regarded as human computer collaboration [38] and theories of collaboration can help to provide new mechanisms for supporting human collaboration. In the field of artificial intelligence, there are numerous attempts to build multiagent systems that can be said to collaborate [23, 22, 40]. For example, Jennings and Mamdani [23] have shown that the present analysis can be used to specify and guide the design of a multiagent system for electric power managment, in which grid diagnostic agents share a joint intention. As a group, the jointly committed agents waste less time in ....

....collaboration [38] and theories of collaboration can help to provide new mechanisms for supporting human collaboration. In the field of artificial intelligence, there are numerous attempts to build multiagent systems that can be said to collaborate [23, 22, 40] For example, Jennings and Mamdani [23] have shown that the present analysis can be used to specify and guide the design of a multiagent system for electric power managment, in which grid diagnostic agents share a joint intention. As a group, the jointly committed agents waste less time in resolving circuit outages than a collection of ....

N. R. Jennings and E. H. Mamdani. Using joint responsibility to coordinate collaborative problem solving in dynamic environments. In Proceedings of the Tenth National Conference on Artificial Intelligence, AAAI Press/MIT Press, Menlo Park, California, July 1992, pp. 269--275.

.... ] The problem of modelling the activity of teams of artificial agents [ Grosz and Kraus, 1993; Kinny et al. 1992 ] is a combination of two sub problems: the first is the modelling of the team itself [ Tidhar, 1993; Werner, 1990 ] and the second is the modelling of the team activity [ Fox, 1981; Jennings and Mamdani, 1992; Kinny et al. 1992 ] The problem of modelling teams and team behavior includes a wide range of subproblems. Such sub problems include the problem of representing teams with a variety of organizational structures and providing a representation of team behavior that allows the implementation of ....

....processes are implicit in the team tactics and that each sub team is responsible to inform other sub team of any change to the tactics. This type of behaviour corresponds to the notions of commitment cf. Cohen and Levesque [ Cohen and Levesque, 1991 ] and responsibility cf. Jennings and Mamdani [ Jennings and Mamdani, 1992 ] 3.2 Implementation In dMARS the agent s beliefs are implemented as relations (or predicates) in a relational database. Since each sub team in a team has at least one role the team itself can be represented as a relation between the team, the sub teams, and the role that the sub team has been ....

N. R. Jennings and E. H. Mamdani. Using joint responsibility to coordinate collaborative problem solving in dynamic environments. In Proceedings of the Tenth National Conference on Artificial Intelligence, pages 269--275, San Jose, CA, USA, 1992.

.... based on this analysis are beginning to be developed (Heeman Hirst, 1992; Edmonds, 1993) Theoretical models of joint action (Cohen Levesque, 240 Chapter 6: Discourse and Dialogue 1991b; Cohen Levesque, 1991a) have been shown to minimize the overall team effort in dynamic, uncertain worlds (Jennings Mamdani, 1992). Thus, if a more general theory of joint action can be applied to dialogue as a special case, an explanation for numerous dialogue phenomena, such as collaboration on reference, confirmations, etc. will be derivable. Furthermore, such a theory offers the possibility for providing a specification ....

Jennings, N. R. and Mamdani, E. H. (1992). Using joint responsibility to coordinate collaborative problem solving in dynamic environments. In Proceedings of the Tenth National Conference on Artificial Intelligence, pages 269--275, Menlo Park, California. American Association for Artificial Intelligence, AAAI Press/MIT Press.

....process. Idle nodes may broadcast node availability announcements. Languages and Applications A prototype of YAMS has been implemented in YAPS (Yet Another Production System) 1] and Franz Lisp and runs on a VAX 11 785 and 2 Sun Workstations which communicate over Ethernet. 2. 2 GRATE GRATE [21] [22] is described as a general purpose cooperation framework based on the idea of Joint Responsibility. It is an expanded version of the earlier GRATE system [21] The system is intended to produce robust industrial applications for use in dynamic and unpredictable environments. Form of an Agent An ....

N.R. Jennings and E.H. Mamdani. Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments. In Proceedings of the Tenth National Conference on Artificial Intelligence, pages 269--275, 1992.

.... but did not inform the AAA or the BAI so that the results they produce were worthless) and distractions from unforeseen events or agents in the community (synchronisation of activities between the AAA and the BAI were affected, causing one agent to needlessly wait for information from the other) [41]. After careful analysis, these problems were attributed to two main factors: i) many of the cooperative actions were not explicitly represented (section 2.1) ii) even when explicit social interactions were set up, the underlying model of cooperation was obscured by a myriad of crucial but ....

N. R. Jennings and E. Mamdani, Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments, in Proc. 10th National Conf. on AI, San Jose, USA (1992) 269-275.

No context found.

Jennings,N.R. and Mamdani,E.H. (1992), "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments." Proc. AAAI-92, 269-75.

....foul ups and inconsistency [12] Responsibility offers a step towards this; providing mechanisms for controlling activity in dynamic and unpredictable environments, whilst retaining a degree of generality and predictability. Empirical evidence to substantiate this claim has been obtained [13]. Compared with groups of selfish problem solvers and communities in which social interactions just emerge, agents organised using the responsibility model performed over twice as well as the other two if there was a greater than 10 chance of the problem solving running into difficulty. ....

N.Jennings & A.Mamdani, (1992), "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments", AAAI, San Jose. (in press)

....when the joint action should terminate. As agents are often situated in evolving and unpredictable environments, and also because they have to take decisions using partial and imprecise information, it is important that the model specifies how to behave in exceptional as well as nominal situations [13]. In particular the model should specify how joint actions may become unstuck, how any problems in the collaboration can be repaired and how agents should respond to these problems both in their local activities and with respect to their acquaintances. In addition, the model should also have a ....

....locally or that some assistance is required (e.g. sharing of information and processing capabilities) but not sufficient to warrant a full scale joint action. The rationale for the last option is that it may not be worth incurring the overhead associated with a distributed planning approach [13] in order to satisfy a small number of interdependencies. In this case assistance will be asked for as and when it is required. Once the need for joint action has been ascertained, the responsibility model requires the following conditions to be fulfilled before a social can commence; i) other ....

[Article contains additional citation context not shown here]

N. R. Jennings and E. H. Mamdani, Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments, Proc. of Tenth National Conference on Artificial Intelligence, San Jose, USA (1992) 269-275

....2 unconstrained as to whether it will use G 2 p,1,3 or G 2 p,1,4 . 3.1. 2 Joint Commitments When agents decide to pursue a joint action, they must jointly commit themselves to a joint goal which will bring about the desired state of affairs (Cohen and Levesque, 1991; Grosz and Sidner, 1990; Jennings, 1992; Kinny et al. 1992; Rao et al. 1992; Searle, 1990; Tuomela and Miller, 1988) This joint commitment has all the aforementioned properties of individual commitment, but it has the additional constraint that it involves more than one agent 3 . This means the overall state of the joint ....

....However in rapidly changing and uncertain contexts, the utility of a relatively sophisticated convention is significantly increased. Indeed empirical evaluation has shown that in such circumstances conventions play a pivotal role in ensuring the community acts in a coherent manner (Jennings and Mamdani, 1992). The lists of both the situations under which commitments should be reassessed and the actions which should be taken in such circumstances can be empty. So an agent can remain permanently committed to a goal even if it has been achieved or an agent can take no action as a result of changes in its ....

Jennings, N. R., and Mamdani, E. H., (1992) "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments" Proc. 10th National Conf. on AI, San Jose, CA., pp 269-275.

.... applications (Jennings and Wittig, 1992; Neches et al. 1991) and providing a more natural representation of distributed problems 1 (eg sensor networks (Lesser and Corkill, 1983) air traffic control (Cammarata et al. 1983) information retrieval (Huhns et al. 1988) and electricity networks (Jennings et al. 1992)) Other potential advantages include: reusability of problem solving components by incorporating the same system into several cooperating communities, an increased set or scope of achievable tasks by sharing resources, improved system robustness by undertaking duplicate tasks using different ....

....convention which describes how to react in such circumstances. Also by opting for dynamic logic to combine plan operators, the expressiveness of the planning language is limited. A general purpose goal interdependence operator would remove this shortcoming. 20 The model of Joint Responsibility (Jennings, 1992) also describes joint action in terms of teams of agents. This model extends Cohen and Levesque s work on joint persistent goals (see 2.2.1) and stipulates that agents should make joint commitments to agreed sequences of actions as well as to the shared objective itself. Commitment to the common ....

[Article contains additional citation context not shown here]

Jennings, N R and Mamdani, E H, 1992. "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments", Proc of 10th National Conf. on Artificial Intelligence, San Jose, USA, pp 269-275.

No context found.

Jennings,N.R. & Mamdani,E.H., (1992), "Using Joint Responsibility to Coordinate Collaborative Problem Solving in Dynamic Environments", AAAI, San Jose, California.

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