Wagner, Th.; Benyo, B.; Lesser, V.; Xuan, P.; Investigating Interactions between Agent Conversations and Agent Control Components; in: [7], pp. 314 -- 330  Home/Search   Document Details and Download   Summary   TOC   Related Articles   Check

....(represented by an input symbol) from outside the modeled system. Models using FSMs or extensions of them where presented by Barbuceanu and Fox (FSM) 1] Barbuceanu and Lo (FSM) 2] von Martial (FSM) 15] Martin et al. PDT) 16] Nodine and Unruh (FSM) 18] and Wagner et al. extended FSM) [24]. Comparing these approaches, it can be seen that all approaches modeling conversations from the viewpoint of an observer are using either STD or petri nets, whereas the approaches using FSM (or state charts) are representing the conversation from the viewpoint of a participating agent. For ....

Wagner, Th.; Benyo, B.; Lesser, V.; Xuan, P.; Investigating Interactions between Agent Conversations and Agent Control Components; in: [7], pp. 314 -- 330

....component (to enable agents to communicate together) etc. We im plemented a coordination component, which provides basic functionality for coordination mechanisms through a com mon substrate. Coordination mechanisms themselves are implemented as Soar rules [6] Recent work, namely GPGP2 [12], implemented this substrate as a Finite State Machine engine. Coordination mechanisms were defined in terms of finite machine scripts that is parsed and compiled and linked to the substrate java code. We took a completely different approach. Our coordination substrate encapsulates a Soar [6] ....

T. Wagner and et al. Investigating Interactions Between Agent Conversations and Agent Control Components. Technical Report 1999-7, University of Massachusetts, 1999.

....a Petri Net representation. Thus a conversation is a combination of protocols being instantiated and manipulated by a particular policy. 4. 1 Conversation modelling A number of modelling techniques have been employed to keep track of agent conversations, including Deterministic Finite Automata [13,16], Enhanced Dooley Graphs [17] and extended UML [18] We use coloured Petri nets [19,20] because their formal properties facilitate the modelling of concurrent conversations and policies in an integrated fashion. Coloured Petri nets are similar to ordinary Petri nets in that they comprise a ....

T. Wagner, B. Benyo, V. Lesser, and P. Xuan. Investigating interactions between agent conversations and agent control components. In Agents 99 Workshop on Conversation Policies. 1999.

....automata to engage in the same conversation. Other conversation models have been developed, using various approaches. Extended FSM models, which, like COOL, focus more on expressivity than adherence to a model, include Kuwabara et al. 21] who add inheritance to conversations, Wagner et al. [39], and Elio and Haddadi [13] who defines a multi level state machine, or Abstract Task Model (ATM) A few others have chosen to stay within the bounds of a DFA, such as Chauhan [7] who uses COOL as the basis for her multi agent development system 13 , Nodine and Unruh [32] and Pitt and Mamdani ....

Thomas Wagner, Brett Benyo, Victor Lesser, and Ping Xuan. Investigating interactions between agent conversations and agent control components. In Working Notes of the Workshop on Specifying and Implementing Conversation Policies,pages 79--88, Seattle, Washington, May 1999.

....reasoning about, and sharing of conversations. Other conversation models have been developed, using various approaches. Extended FSM models, which, likeCOOL, focus moreon expressivitythan adherencetoamodel,includeKuwabara et al. 27] who add inheritance to conversations,Wagner et al. [44], and Elio and Haddadi [13] who de nes a multi level state machine, or Abstract Task Model (ATM) A few others have chosen to stay within the bounds of a DFA, suchas Chauhan [6] who uses COOL as the basis for her multi agent development system # , Nodine and Unruh [38] and Pitt and Mamdani ....

Thomas Wagner, Brett Benyo, Victor Lesser, and Ping Xuan. Investigating interactions between agent conversations and agent control components. In Working Notes of the Workshop on Specifying and Implementing Conversation Policies,pages 79-88, Seattle, Washington, May 1999.

....extensive work being done in the field of agent conversation modelling. The suitability of Petri Nets as a compact and uniform model has been already proposed [4, For reference and comparison we have chosen some other traditional solutions based on the Deterministic Finite Automata (DFAs) [2,24,14], Enhanced Dooley Graphs [18] and extended UML [16, 17] formalisms. Deterministic Finite Automata represent the simplest and most straightforward modelling formalism used for conversations. They are suitable for specifying all the states the conversation must go through, and can be used to ....

T. Wagner, B. Benyo, V. Lesser, and P. Xuan. Investigating interactions between agent conversations and agent control components. In Agents 99 Workshop on Conversation Policies. 1999.

....or too computationally expensive to implement. Other conversation models have been developed, using various approaches. Extended FSM models, which, like COOL, focus more on expressivity than adherence to a model include Kuwabara et al. 22, 21] who add inheritance to conversations; Wagner et al. [46]; and Elio and Haddadi [14] who defines a multilevel state machine, or ATM. A few others have chosen to stay within the bounds of a DFA, such as Chauhan [7] who uses COOL as the basis for her multi agent development system, 1 Nodine and Unruh [34, 35] who use conversation specifications to ....

T. Wagner, B. Benyo, V. Lesser, and P. Xuan. Investigating interactions between agent conversations and agent control components. In Working Notes of the Workshop on Specifying and Implementing Conversation Policies, pages 79--88, Seattle, Washington, May 1999.

....reasoning about and sharing of conversations. Other conversation models have been developed, using various approaches. Extended FSM models, which, like COOL, focus more on expressivity than adherence to a model, include Kuwabara et al. 29] who add inheritance to conversations, Wagner et al. [43], and Elio and Haddadi [13] who defines a multi level state machine, or Abstract Task Model (ATM) A few others have chosen to stay within the bounds of a DFA, such as Chauhan [6] who uses COOL as the basis for her multi agent development system 1 , Nodine and Unruh [37] and Pitt and Mamdani ....

Thomas Wagner, Brett Benyo, Victor Lesser, and Ping Xuan. Investigating interactions between agent conversations and agent control components. In Working Notes of the Workshop on Specifying and Implementing Conversation Policies, pages 79--88, Seattle, Washington, May 1999.

....fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and or a fee. Copyright 2000 ACM 0 89791 88 6 97 05 . 5.00 it via a model of the agent s problem solving process. The modeling framework we use is called TMS [6, 18] TMS task structures resemble complex and or graphs or HTNs, i.e. TMS is a hierarchical decomposition framework. Notable features of TMS models include the ....

....and resource constraints. The core of this ability is the Design to Criteria [29, 27, 26, 23] agent scheduler it is the TMS analysis expert. When multi agent systems are constructed, a coordination module is added to the agent and often the module is GPGP [7, 17] or one of its descendents [25] Though we have used proprietary components for this as well [16, 11] A natural question is where do the TMS models come from In our work TMS task structures are often hand crafted though the general idea is that a domain expert, e.g. a blackboard problem solver or a planner, should ....

Thomas Wagner, Brett Benyo, Victor Lesser, and Ping Xuan. Investigating Interactions Between Agent Conversations and Agent Control Components. In Frank Dignum and Mark Greaves, editors, Issues in Agent Communication, Lecture Notes in Artificial Intelligence, pages 314--331. Springer-Verlag, Berlin, 2000.

....provides the agent with ability to reason about the trade offs of different possible activities while respecting the criteria requirement. The GPGP coordination module is a domain independent coordination framework that is based on Decker s original GPGP ( 5] and our recent extensions (GPGP2) [12]. It communicates with other agents, it also communicates with DTC scheduler, but these two kinds of communication are transparent to the domain problem solver. The main functions of the GPGP system include: ffl Exchange Non Local Viewpoints Agents exchange information about the activities that ....

Thomas Wagner, Brett Benyo, Victor Lesser, and Ping Xuan. Investigating Interactions Between Agent Conversations and Agent Control Components. In Agents 99 Workshop on Conversation Policies, 1999.

....and capable of expressing complex task environments. In terms of reasoning and coordination using the T MS, our discussion will focus around the scheduling framework of Design to Criteria scheduling [17] and the Generalized Partial Global Planning (GPGP GPGP2) family of coordination mechanisms [3, 15]. The basic building blocks in T MS are tasks, methods, and interrelationships. Figure 1 shows (partial) specifications of a task, a method, and an enables interrelationship. spec task (spec method (label tB) label m2) supertasks tA) supertasks tC2) subtasks tC1 tC2) outcomes (qaf q min) o1 ....

Thomas Wagner, Brett Benyo, Victor Lesser, and Ping Xuan. Investigating Interactions Between Agent Conversations and Agent Control Components. In Agents 99 Workshop on Conversation Policies, 1999. Also available as UMASS CS TR 99-07.

....provides the agent with ability to reason about the trade offs of different possible activities while respecting the criteria requirement. The GPGP coordination module is a domain independent coordination framework that is based on Decker s original GPGP ( 5] and our recent extensions (GPGP2) [12]. It communicates with other agents, it also communicates with DTC scheduler, but these two kinds of communication are transparent to the domain problem solver. The main functions of the GPGP system include: Exchange Non Local Viewpoints Agents exchange information about the activities that ....

Thomas Wagner, Brett Benyo, Victor Lesser, and Ping Xuan. Investigating Interactions Between Agent Conversations and Agent Control Components. In Agents 99 Workshop on Conversation Policies, 1999.

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