Kinny, David, and Michael Georgeff, "Modelling and Design of Multi-Agent Systems," Intelligent Agents III: Proceedings of the Third International workshop on Agent Theories, Architectures, and Languages (ATAL'96), ed., Springer, Heidelberg, 1996.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....able to model real complex systems with clearly distributed characteristics. These works are principally based on a view of the system as a computational organization which consist of di#erent interacting entities. Some of the relevant existing approaches are: the work of Kinny on the BDI model [22], the GAIA methodology [30] the works of Kendall [21] and Bursmeister [6] the AUML approach [24] the specification of DESIRE [5] the methodologies MAS CommonKADS [15] and MASSIVE [23] and more recently the works of TROPOS [26] MaSE [29] and Message [10] The majority of these approaches ....

D. Kinny and M. George#. Modelling and design of multi-agent systems. Technical Report 59, Australian Artificial Intelligence Institute, Melbourne, Australia, 1996.

....and capabilities: what other components they interact with, and via which event database; and for events and databases: what components read post write receive them. 4 Related Work There is currently a large amount of work being done in agent oriented software engineering methodologies (e.g. [1, 3, 4, 5, 6, 9, 10, 11, 12, 13, 14, 18, 19, 22], and see the references in [15] It is not possible in a brief conference paper to do justice to the significant amount of work being done. The GAIA methodology [22] has, like Prometheus, been developed over a number of years by people experienced in building agent systems. However we found ....

D. Kinny and M. Georgefl. Modelling and design of multiagent systems. In Intelligent Agents III.' Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL- 96). LNAI 1193. SpringerVerlag, 1996.

....The diamond symbol with an indicates an exclusive or decision. If a proposal is o#ered, Customer has a choice of either accepting or canceling the proposal. The internal processing of Shopping Cart s checkout plan is described in Figure 16. At the lowest level, we use plan diagrams [31], to specify the internal processing of atomic actors. Each identified plan is specified as a plan diagram, which is denoted by a rectangular box. The lower section, the plan graph, is a state transition diagram. However, plan graphs are not just descriptions of system behavior developed during ....

....them and that we do not necessarily operationalize these intentional and social structures early on during the development process, avoiding to freeze solutions to a given requirement in the produced software designs. There already exist some proposals for agent oriented software development like [3,27,28,31,41,46,48]. Such proposals are mostly extensions to known objectoriented and or knowledge engineering methodologies. Moreover, all these proposals focus on design as opposed to requirements analysis and are therefore considerably narrower in scope than Tropos. Indeed, Tropos proposes to 28 adopt ....

D. Kinny and M. George#. Modelling and design of multi-agent systems. In Proc. of the 3rd Int. Workshop on Intelligent Agents: Agent Theories, Architectures, and Languages, ATAL'96, pages 1--20, Budapest, Hungary, Aug. 1997.

....and systems for designing MAS and developing agent oriented application are still quite rare, even if more and more researchers have begun to work in this field in the last period. About design methodologies the research work is really in its infancy. We can cite a paper by Kinny and Georgeff [KG96] regarding a methodology for the design and specification of large scale commercial and industrial software systems based on MAS (modelled extending object oriented techniques) and a book by Muller [Mul96] perhaps the first monograph about theory and practice in designing MAS. Regarding systems ....

D. Kinny and M. Georgeff. Modelling and Design of Multi--Agent Systems. In J. P. Muller, M. Wooldridge, and N. R. Jennings, editors, Intelligent Agents III, pages 1--20. Springer-Verlag, 1996.

....agent should behave that is independent of the specific agent architecture employed, we can demonstrate its value by closely marrying it with equally intuitive design and implementation methods. The need for intuitive modelling techniques in agent oriented programming has been noted elsewhere [4]. In this section we present an agent architecture specifically developed for autonomous agents with the kinds of requirements discussed in Section 2. The architecture is inspired by the distributed Multi Agent Reasoning System (dMARS) 3] a highly successful commercial architecture based on the ....

David Kinny and Michael Georgeff. Modelling and design of multi-agent systems. In Jorg P. Muller, Michael J Wooldridge, and Nicholas R. Jennings, editors, Intelligent Agents III: Proceedings of the ECAI'96 Workshop on Agent Theories, Architectures, and Languages, LNAI 1193, pages 1--20. Springer-Verlag, 1996.

....engineering are: i) an understanding of the situations in which agent solutions are appropriate; and (ii) principled but informal development techniques for agent systems. While some attention has been given to the latter (in the form of analysis and design methodologies for agent systems [30, 60, 39, 10, 3, 29, 53]) almost no attention has been given to the former (but see [59] 8. How to Find Out More About Agents There are now many introductions to intelligent agents and multiagent systems. Ferber [14] is an undergraduate textbook, although as its name suggests, this volume focussed on multiagent ....

D. Kinny and M. George#. Modelling and design of multi-agent systems. In J. P. Muller, M. Wooldridge, and N. R. Jennings, editors, Intelligent Agents III (LNAI Volume 1193), pages 1--20. Springer-Verlag: Berlin, Germany, 1997.

....focuses on the protocol between Customer and Shopping Cart which consists of a customization of the Contract Net FIPA agent pattern [21] Such a protocol describes a communication pattern among actors, as well as constraints on the contents of the messages they exchange. We use plan diagrams [18], based on state charts and activity diagrams, to specify the internal processing (tasks) of atomic actors. The initial transition of the plan diagram is labeled with an activation event (Press checkout button) and activation condition ( checkout button activated] which determine when and in what ....

....in addition to be systems requiring flexibility and dynamicity, they are built on mental states like beliefs, intentions, desires or commitments and considered societies of software entities. On the other hand, there already exist some proposals for agent oriented software development like [10, 16, 17, 18, 25]. Such proposals are mostly extensions to known object oriented and or knowledge engineering methodologies. Moreover, all these proposals focus on design as opposed to requirements analysis and are therefore considerably narrower in scope than Tropos. Indeed, Tropos proposes to adopt the ....

Kinny, D. and Georgeff, M., "Modelling and Design of Multi-Agent System", Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL'96), pp. 1-20, Budapest, Hungary, August 1996.

....failure FIPA Contract Net Protocol accept proposal, succeeded, failure propose, cancel proposal checkout rfp, refuse, not understood, 12 19 00 at 9:31 53 Customer, Shopping Cart Fig. 8. Agent interaction protocol focusing on a checkout dialogue At the lowest level, we use plan diagrams [12] (See Figure 9) to specify the internal processing of atomic actors. The initial transition of the plan diagram is labeled with an activation event (Press checkout button) and activation condition ( checkout button activated] which determine when and in what context the plan should be activated. ....

Kinny, D. and Georgeff, M., "Modelling and Design of Multi-Agent System", Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL-96), Budapest, Hungary, August 1996, pp. 1-20.

....The diamond symbol with an indicates an exclusive or decision. If a proposal is o#ered, Customer has a choice of either accepting or canceling the proposal. The internal processing of Shopping Cart s checkout plan is described in Figure 18. At the lowest level, we use plan diagrams [30], to specify the internal processing of atomic actors. Each identified plan is specified as a plan diagram, which is denoted by a rectangular box. The lower section, the plan graph, is a state transition diagram. However, plan graphs are not just descriptions of system behavior developed during ....

....them and that we do not necessarily operationalize these intentional and social structures early on during the development process, avoiding to freeze solutions to a given requirement in the produced software designs. There already exist some proposals for agent oriented software development like [3,26,27,30,41,46,48]. Such proposals are mostly extensions to known objectoriented and or knowledge engineering methodologies. Moreover, all these proposals focus on design as opposed to requirements analysis and are therefore considerably narrower in scope than Tropos. Indeed, Tropos proposes to adopt the ....

D. Kinny and M. George#. Modelling and design of multi-agent systems. In Proc. of the 3th Int. Workshop on Intelligent Agents: Agent Theories, Architectures, and Languages, ATAL'96, pages 1--20, Budapest, Hungary, Aug. 1997.

....for the next generation of mainstream software systems. However, it is essential to have complexity management and design methodology to support the development of large embedded systems. Lack of them is the main reason why many actual projects trying to apply agent technique ended with failure [5, 6, 8]. The importance of agent oriented methodology can t be more emphasized, and is becoming the most promising research area in agent eld. Much work has been done in this area, from formal speci cation[1, 2] to analysis and design[3, 4, 5, 9] which can be classi ed into two types. The rst kind is ....

....projects trying to apply agent technique ended with failure [5, 6, 8] The importance of agent oriented methodology can t be more emphasized, and is becoming the most promising research area in agent eld. Much work has been done in this area, from formal speci cation[1, 2] to analysis and design[3, 4, 5, 9], which can be classi ed into two types. The rst kind is component based methodology, where agent is modeled as a composite component from lower level components. The other kind is BDIbased, which combines BDI architecture with object oriented technique and has been shown to be a feasible ....

D. Kinny and M. George. Modelling and design of multiagent systems. In J. P. Muller, M. Wooldridge, and N. R. Jennings, eds, Intelligent Agents III (LNAI Volume 1193), pages 1-20. Springer-Verlag: Berlin, Germany, 1997.

.... processOrder(this) profile updated [i 3] CC# not valid] CC# valid] confirm( confirm( Press confirm button any [ Cancel button pressed] OR [timeout 90 sec] shoppingCart.initialize( Figure 10: A plan diagram for checkout At the lowest level, we use plan diagrams [12], to specify the internal processing of atomic actors. The initial transition of the plan diagram is labeled with an activation event (Press checkout button) and activation condition ( checkout button activated] which determine when and in what context the plan should be activated. Transitions ....

Kinny, D. and Georgeff, M., "Modelling and Design of Multi-Agent System", Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL-96), Budapest, Hungary, August 1996, pp. 1-20.

....pressed] OR [timeout 90 sec] shoppingCart.initialize( Figure 11: A plan diagram for checking out At the lowest level, state charts and activity diagrams can be used to specify the internal processing (plans) of actors who are no more aggregates. We use them as plan diagrams as proposed in [Kin96,Kin96a]. Each identified plan is specified as a plan diagram, which is denoted by a rectangular box. The lower section, the plan graph, is a state transition diagram. However, plan graphs are not just descriptions of system behavior developed during analysis. Rather, they are directly executable ....

Kinny, D. and Georgeff, M., "Modelling and Design of Multi-Agent System", Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL-96), Budapest, Hungary, August 1996, pp. 1-20.

.... is to display a rational behaviour, it needs to have a module capable of handling its beliefs (see e.g. the Procedural Reasoning System (PRS) proposed in [GeIn89] the BDI (Belief Desire Intention) architecture considered in [Brat87] BIP88] CoLe90] RaGe91] RaGe93] RaGe95a] RaGe95b] [KiGe97] or [Morl97] the implicit agent architecture developed in [Denn78] and [Denn84] the Agent oriented programming paradigm proposed in [Shoh90] or [Shoh93] or the defeasible reasoner suggested in [Poll90] or [Poll95] In this document we take the expression rational agents to refer to those ....

Kinny, D., Georgeff, M., "Modelling and design of multi-agent systems", in

....how it interfaces with the production process. 2 Related Work With the shift from laboratory to industrial applications, it has become increasingly apparent that existing methodologies, such as purely object oriented approaches, are insufficient to capture the key features of agent based systems [1,14]. This experience has led to the development of distinctively agent oriented design methodologies over the last few years. Most agent oriented methodologies have been extensions of existing methodologies, in particular knowledge oriented and object oriented approaches. Only recently have ....

....possible collaborators, required expertise, and co operation mechanisms used. The most important advantage of the concept of a role is that it can be freely assign and reassigned to agents, as long as the agent assigned to the role fulfils the role s requirements. Role based methodologies, e.g. [5,14,15,17,18,23], use this abstraction to create a model of system behaviour, and then identify agents by mapping the roles to agent instances. The Gaia methodology [23] for instance, aggregates roles into agent types and instantiates as many agents as necessary in a given scenario. Most role based ....

D. Kinny, M. Georgeff: "Modelling and Design of Multi-Agent Systems". In J.P. Mller, M.J. Wooldridge, N.R. Jennings (eds.), Intelligent Agents III (ATAL'96), LNAI 1193, pages 1 -- 20. Springer-Verlag: Berlin, Germany, 1997.

....the actor, i.e. the aircraft controls, are extremely complex and allow many degrees of freedom [ Tambe et al. 1995a ] 1.3 Related Work Recently there has been a lot of interest in development methodologies for agents. Examples include methodologies for Belief Desire Intention (BDI) agents [ Kinny and George , 1996 ] for behavior based agents [ Bryson, 1998 ] for distributed multi agent systems [ Bussman, 1998 ] for safety critical multi agent systems [ Bouchefra et al. 1998 ] and a more general method for a range of agent oriented systems [ Wooldridge et al. 1999 ] Development environments, often ....

David Kinny and Michael George. Modelling and design of multiagent systems. In Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages. Lecture Notes in Computer Science, 1996.

....the specification into an executable form, directly execute the abstract specification, automatically translate the specification into an executable form. The first approach is already hard for reactive systems and even more for multi agent systems, nevertheless some particular examples [LD95, KG96] have been realized. Directly executing an agent specification (expressed as a logic formula OE) means both generating a model M for OE (if OE represents some property of the agent) or giving a proof of OE (if OE encodes the behaviour of the agent, as in a logic programming setting) Miller ....

....and systems for designing MAS and developing agent oriented application are still quite rare, even if more and more researchers have begun to work in this field in the last period. About design methodologies the research work is really in its infancy. We can cite a paper by Kinny and Georgeff [KG96] regarding a methodology for the design and specification of large scale commercial and industrial software systems based on MAS (modelled extending object oriented techniques) and a book by Muller [Mul96] perhaps the first monograph about theory and practice in designing MAS. Regarding systems ....

D. Kinny and M. Georgeff. Modelling and Design of Multi--Agent Systems. In J. P. Muller, M. Wooldridge, and N. R. Jennings, editors, Intelligent Agents III, pages 1--20. Springer-Verlag, 1996.

....engineering are: i) an understanding of the situations in which agent solutions are appropriate; and (ii) principled but informal development techniques for agent systems. While some 22 attention has been given to the latter (in the form of analysis and design methodologies for agent systems [21]) almost no attention has been given to the former (but see [43] ....

D. Kinny and M. Georgeff (1997) "Modelling and design of multi-agent systems" In J. P. Mueller, M. Wooldridge, and N. R. Jennings, editors, Intelligent Agents III (LNAI Volume 1193), 1-20. Springer-Verlag.

....The idea of maintenance also emerges in implemented planning systems, which from the time of Waldinger [30] have considered maintenance as a key functionality in constructing effective plans, especially when more than one goal must be brought about. Maintenance features also in the dMARS system [15]. It is therefore quite interesting that maintenance has not received corresponding attention in the formal reasoning about action community. Consequently, implemented systems usually handle maintenance in a seat of the pants manner. Previous TL approaches assume that all changes in the ....

David Kinny and Michael P. Georgeff. Modelling and design of multi-agent systems. In Intelligent Agents III: Agent Theories, Architectures, and Languages, pages 1--20, 1997.

....and short time objectives. The desire for an efficient use of the multi agent technology, i.e. software agents [17] has also made research on methodologies very popular. The functional analysis within a methodology, like CoMoMAS, demonstrate the closeness of multiagent and software engineering [14]. Functional models are an interesting starting point for development since they allow us to identify clearly control and data dependencies; this is valuable information which can be used to decide the distribution of tasks between agents and human developers. The advocated functional analysis of ....

D. Kinny and M. Georgeff. Modelling and design of multiagent system. In ECAI-96 workshop on ATAL, pages 1--20. LNAI 1193, 1996.

....be reused [8] Such a design decision holds information about how to proceed for transforming one model template into another one; they can also be used as index for model selection from a library. Noteworthy are also object oriented approaches describing the functional decomposition of a system. [12] views a MAS from an external and an internal viewpoint describing the interaction between abstract and concrete agent classes obtained from system decomposition and the functional decomposition of agents. 14] introduces a hierarchy of entities comprising autonomous agents, agents and objects as ....

D. Kinny and M. Georgeff. Modelling and design of multi-agent system. In ECAI-96 workshop on ATAL, pages 1--20, Budapest, Hungary, 1996. LNAI Series Vol. 1193.

....such as [5] However, it is beyond the scope of this article to compare the Gaia definition and use of these concepts with this heritage. Instead, we will focus here on the relationship between Gaia and other approaches based that build upon OO techniques, in particular the KGR approach [23, 24]. But it is perhaps useful to begin by summarizing why OO modelling techniques and design methodologies themselves are not directly applicable to multi agent system design. 6.1. Shortcomings of object oriented techniques The first problem concerns the modelling of individual agents or agent ....

....differently in an agent oriented approach. The uniformity and concreteness of the object model is the basis of the problem; OO methodologies provide guidance or inspiration rather than a directly useful approach to analysis and design. 6.2. Comparison with the KGR approach The KGR approach [23, 24] was developed to fulfill the need for a principled approach to the specification of complex multi agent systems based on the beliefdesire intention (BDI) technology of the Procedural Reasoning System (prs) and the Distributed Multi Agent Reasoning System (dmars) 8, 25] A key motivation of the ....

D. Kinny and M. Georgeff, "Modelling and design of multi-agent systems," in Intelligent Agents III (LNAI Vol. 1193),J.P.M uller, M. Wooldridge, and N. R. Jennings (Eds.), Springer-Verlag: Berlin, Germany, 1997, pp. 1--20.

....such as [5] However, it is beyond the scope of this article to compare the Gaia definition and use of these concepts with this heritage. Instead, we will focus here on the relationship between Gaia and other approaches based that build upon OO techniques, in particular the kgr approach [24, 23]. But it is perhaps useful to begin by summarizing why OO modelling techniques and design methodologies themselves are not directly applicable to multi agent system design. 6.1. Shortcomings of Object Oriented techniques The first problem concerns the modelling of individual agents or agent ....

....differently in an agent oriented approach. The uniformity and concreteness of the object model is the basis of the problem; OO methodologies provide guidance or inspiration rather than a directly useful approach to analysis and design. 6.2. Comparison with the KGR approach The KGR approach [24, 23] was developed to fulfill the need for a principled approach to the specification of complex multi agent systems based on the belief desire intention (BDI) technology of the Procedural Reasoning System (PRS) and the Distributed Multi Agent Reasoning System (DMARS) 25, 8] A key motivation of the ....

D. Kinny and M. Georgeff. Modelling and design of multi-agent systems. In J. P. Muller, M. Wooldridge, and N. R. Jennings, editors, Intelligent Agents III (LNAI Volume 1193), pages 1--20. Springer-Verlag: Berlin, Germany, 1997.

No context found.

Kinny, David, and Michael Georgeff, "Modelling and Design of Multi-Agent Systems," Intelligent Agents III: Proceedings of the Third International workshop on Agent Theories, Architectures, and Languages (ATAL'96), ed., Springer, Heidelberg, 1996.

No context found.

D. Kinny and M. Georgeff. Modelling and design of multi-agent systems. In J. P. Muller, M. Wooldridge, and N. R. Jennings, editors, Intelligent Agents III (LNAI Volume 1193), pages 1--20. Springer-Verlag: Berlin, Germany, 1997.

No context found.

D. Kinny and M. Georgeff. Modelling and design of multi-agent systems. In Intelligent Agents III: Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL-96). LNAI 1193. Springer-Verlag, 1996.

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