[Show abstract][Hide abstract] ABSTRACT: Le paradigme des systèmes multi-agents (SMA) est approprié pour des applications distribuées sans contrôle centralisé et pour lesquelles il est nécessaire qu'un sousensemble des agents collabore afin de résoudre un problème global. Les systèmes multiagents comme toute application distribuée, asynchrone et à faible couplage sont difficiles à concevoir et à développer. Nous proposons de faciliter leur conception par la réutilisation de patterns logiciels. Les patterns constituent des solutions génériques à des problèmes fréquemment rencontrés. Nous avons conçu un métamodèle représentant et structurant les concepts inhérents aux SMA. A partir de ce modèle, douze patterns d'analyse décrivant les éléments conceptuels nécessaires à la spécification d'applications orientées agents ont été conçus, ainsi que des patterns de support d'utilisation facilitant la réutilisation de ces patterns lors de la phase d'analyse du processus d'ingénierie des SMA.
[Show abstract][Hide abstract] ABSTRACT: Agent-Oriented Simulations (AOS) are helpful to under- stand real complex systems particularly via their mod- eling ability and the observations they provide. To fully exploit those simulations, AOS platforms should oer shorter simulation time even if simulated systems are more and more complex. In this paper, we propose a method to minimize the ex- ecution time of AOS through the parallel execution of simulation agents. Our solution is ensured by a net- work of AOS platforms that relies on the identication of agents' temporal dependencies. Our proposal is based on the Temporality model that allows the agents to express their temporal dynamics.
[Show abstract][Hide abstract] ABSTRACT: We present MABLE, a fully implemented programming language for multi- agent systems, which is intended to support the automatic verification of such systems via model checking. In addition to the conventional constructs of im- perative programming languages, MABLE provides a number of agent-oriented development features. First, agents in MABLE are endowed with a BDI-like men- tal state: they have data structures corresponding to beliefs, desir es, and inten- tions, and these mental states may be arbitrarily nested. Se cond, agents in MABLE communicate via ACL-like performatives: however, neither the performatives nor their semantics are hardwired into the language. It is possi ble to define the perfor- matives and the semantics of these performatives independently of the system in which they are used. Using this feature, a developer can explore the design space of ACL performatives and semantics without changing the target system. Finally, MABLE supports automatic verification via model checking. Claims about the be- haviour of a MABLE system can be expressed in a linear-time BDI-like logic, and the truth, or otherwise, of these claims can be automaticall y determined. Follow- ing a description of the MABLE language and the language of MABLE claims, we present two case studies to illustrate the language and its u se in the verification of multiagent systems. We then describe the key ideas underpinning the current implementation of MABLE. Finally, we survey related work, and discuss some avenues for future research.
Preview · Article · Apr 2006 · International Journal of Artificial Intelligence Tools
[Show abstract][Hide abstract] ABSTRACT: Multiagent systems are becoming increasingly popular as a new programming paradigm that provides the right abstraction level
and the right model to build a lot of distributed applications. Its basic components are agents which are encapsulated computer systems that are situated in some environment and that are capable of flexible, autonomous action in
that environment in order to meet their design objectives .
[Show abstract][Hide abstract] ABSTRACT: The problem of checking that agents correctly implement the semantics of an agent communication language has become increasingly important as agent technology makes its transition from the research laboratory to field-tested applications. In this paper, we show how model checking techniques can be applied this problem. Model checking is a technique developed within the formal methods community for automatically verifying that finite-state concurrent systems implement temporal logic specifications. We first describe a variation of the MABLE multiagent BDI programming language, which permits the semantics (pre- and post-conditions) of ACL performatives to be defined separately from a system where these semantics are used. We then show how assertions defining compliance to the semantics of an ACL can be captured as claims about MABLE agents, expressed using MABLE's associated assertion language. In this way, compliance to ACL semantics reduces to a conventional model checking problem. We illustrate our approach with a number of short case studies.
[Show abstract][Hide abstract] ABSTRACT: Designing agent interaction protocols need first to consider what the requirements are. This is done in the requirement analysis
phase. The output of this phase is an informal document written in natural language. To our best knowledge, this phase is
barely considered in the literature neither in communication protocol engineering nor in interaction protocol engineering.
As a consequence, it is difficult for designers to do it easily. Experience seems to be the key. In order to help designers,
we propose to structure the requirement analysis document into fields. These fields gather protocol’s features. This paper
presents such document and applies the structuration to the electronic commerce protocol NetBill.
[Show abstract][Hide abstract] ABSTRACT: Several methodologies are supplied to multiagent system designers to help them defining their agents and their multiagent
systems. These methodologies focus mainly on agents and on multiagent systems and barely consider how to design interaction
protocols. A problem could emerge of this lack since interaction protocols are more and more complex. The aim of this article
is to present our proposal of interaction protocol engineering which is based on the communication protocol engineering .
Interaction protocol engineering allows designers to define protocols from scratch. Our proposal is composed of five stages:
analysis, formal description, validation, protocol synthesis and conformance testing.
[Show abstract][Hide abstract] ABSTRACT: In this paper we focus on the analysis and design stages of the protocol engineering development cycle. We start by sketching an application framework dedicated to a Web-based learning environment called Baghera whose aim is to teach geometry problems. We then apply our protocol engineering process to protocols for checking mathematical proofs a student happens to build. The following section discusses the analysis stage of such a protocol. We then briefly introduce our component-based formal specification language in order to then describe the protocol's design stage. Finally we present a tool built upon the FIPA norm (making use of the PDN or UAML notation) which supports the analysis and design of interaction protocols.
[Show abstract][Hide abstract] ABSTRACT: MABLE is a language for the design and automatic verification of multi-agent systems. MABLE is essentially a conventional imperative programming language, enriched by constructs from the agent-oriented programming paradigm. A MABLE system contains a number of agents, programmed using the MABLE imperative programming language. Agents in MABLE have a mental state consisting of beliefs, desires and intentions. Agents communicate using request and inform performatives, in the style of the fipa agent communication language. MABLE systems may be augmented by the addition of formal claims about the system, expressed using a quantified, linear temporal belief-desire-intention logic. MABLE has been fully implemented, and makes use of the spin model checker to automatically verify the truth or falsity of claims.
[Show abstract][Hide abstract] ABSTRACT: Successful development of agent interaction protocols re- quires modeling methods and tools that support a relatively complete
development lifecycle.Agent-based systems are inherently complexbut exhibit many similarities to object-oriented systems.For
these reasons not only current modeling languages need to be extended,but also re-lated tools should be provided for agent
interaction protocol design to be supported.In this paper,we focus on the design stage of an agent in-teraction protocol development
cycle.We start by giving general criteria for comparing agent modeling languages.The ones we take into consid-eration in this
paper are extensions of Agent-UML and FIPA-UAML languages.We describe these languages and discuss some extensions on a simpli
.ed application of the Netbill electronic commerce protocol.We then brie .y introduce a component-based formal speci .cation
language in order to support the protocol ? design stage and present a tool built upon the FIPA norm (making use of the PDN
or UAML notation)which supports the analysis and design of interaction protocols.
[Show abstract][Hide abstract] ABSTRACT: The main formalisms used for designing interaction protocols in multiagent systems do not allow for an easy definition of reusable protocols. In this paper, we propose a protocols description language based on the combination of components called micro-protocols corresponding to sets of performatives which represent pieces of an interaction. We formally introduce these components and present the communication protocol description language we have defined in order to combine these microprotocols into whole interaction protocols. We then demonstrate the ease of use of our approach on the well-known contract net protocol example, and compare our model to similar approaches drawn from the fields of distributed systems and multiagent systems. Prior to concluding on the limits and possible extensions of this approach, we detail a software tool which supports the design and implementation of interaction protocols according to our formalism.
[Show abstract][Hide abstract] ABSTRACT: The design and development of open multiagent sys-tems is one of the main areas in multiagent research. Human soci-eties have successfully coped with a similar issue by creating insti-tutions, which can broadly be understood as formal or semi-formal frameworks that provide commonly understood collections of rules within which people can interoperate. In ongoing work, we have been developing a framework called ISLANDER for the specification and developed of electronic institutions, within which software agents can meet and interoperate within a commonly understood terms of reference. This framework includes both a formal specification lan-guage for institutions and software tools to assist in their construc-tion. In this paper, we are concerned with providing tools to support the design-time verification and validation of such institutions. We present some preliminary results in the use of a SPIN-based multia-gent model checking framework called MABLE to enable verifica-tion of ISLANDER models.