Open Agent Systems Research Articles

Decision making in open agent systems

AbstractIn many real‐world applications of AI, the set of actors and tasks are not constant, but instead change over time. Robots tasked with suppressing wildfires eventually run out of limited suppressant resources and need to temporarily disengage from the collaborative work in order to recharge, or they might become damaged and leave the environment permanently. In a large business organization, objectives and goals change with the market, requiring workers to adapt to perform different sets of tasks across time. We call these multiagent systems (MAS) open agent systems (OASYS), and the openness of the sets of agents and tasks necessitates new capabilities and modeling for decision making compared to planning and learning in closed environments. In this article, we discuss three notions of openness: agent openness, task openness, and type openness. We also review the past and current research on addressing the novel challenges brought about by openness in OASYS. We share lessons learned from these efforts and suggest directions for promising future work in this area. We also encourage the community to engage and participate in this area of MAS research to address critical real‐world problems in the application of AI to enhance our daily lives.

Scalable Decision-Theoretic Planning in Open and Typed Multiagent Systems

In open agent systems, the set of agents that are cooperating or competing changes over time and in ways that are nontrivial to predict. For example, if collaborative robots were tasked with fighting wildfires, they may run out of suppressants and be temporarily unavailable to assist their peers. We consider the problem of planning in these contexts with the additional challenges that the agents are unable to communicate with each other and that there are many of them. Because an agent's optimal action depends on the actions of others, each agent must not only predict the actions of its peers, but, before that, reason whether they are even present to perform an action. Addressing openness thus requires agents to model each other's presence, which becomes computationally intractable with high numbers of agents. We present a novel, principled, and scalable method in this context that enables an agent to reason about others' presence in its shared environment and their actions. Our method extrapolates models of a few peers to the overall behavior of the many-agent system, and combines it with a generalization of Monte Carlo tree search to perform individual agent reasoning in many-agent open environments. Theoretical analyses establish the number of agents to model in order to achieve acceptable worst case bounds on extrapolation error, as well as regret bounds on the agent's utility from modeling only some neighbors. Simulations of multiagent wildfire suppression problems demonstrate our approach's efficacy compared with alternative baselines.

Transition systems for designing and reasoning about norms

The design and analysis of norms is a somewhat neglected topic in AI and Law, but this is not so in other areas of Computer Science. In recent years powerful techniques to model and analyse norms have been developed in the Multi-Agent Systems community, driven both by the practical need to regulate electronic institutions and open agent systems, and by a theoretical interest in mechanism design and normative systems. Agent based techniques often rely heavily on enforcing norms using the software to prevent violation, but I will also discuss the use of sanctions and rewards, and the conditions under which compliance by autonomous agents (including humans) can be expected or encouraged without sanctions or rewards. In the course of the paper a suggested framework for the exploration of these issues is developed.

Dynamic specification of open agent systems

Multi-Agent Systems (MAS) where the agents are developed by different parties and serve different interests, are often classified as ‘open’. The specification of open MAS is largely seen as a design-time activity. Moreover, there is no support for run-time specification modification. Due to environmental, social or other conditions, however, it is often required to revise the MAS specification during its execution. To address this requirement, we present an infrastructure for ‘dynamic’ open MAS specifications, that is, specifications that may be modified at run-time by the agents. We adopt a bird's eye view of an open MAS, as opposed to an agent's own perspective whereby it reasons about how it should act. The infrastructure consists of well-defined procedures for proposing a modification of the ‘rules of the game’, as well as decision-making over and enactment of proposed modifications. We evaluate proposals for rule modification by modelling a dynamic specification as a metric space, and by considering the effects of accepting a proposal on system utility. Furthermore, we constrain the enactment of proposals that do not meet the evaluation criteria. We employ the action language C+ to formalize dynamic specifications, and the ‘Causal Calculator’ implementation of C+ to execute the specifications. We illustrate our infrastructure by presenting a dynamic specification of a resource-sharing protocol.

P.3.c.059 Oxidative stress status in schizophrenic patients treated with typical versus atypical antipsychotics

This volume contains the proceedings of the FCT′99 Workshop on Distributed Systems held on 2-3 September 1999 in Iassy, Romania as a satellite event to FCT′99 — the 12th International Symposium on Fundamentals of Computation Theory.Topics: The declared aim of the workshop was to select papers presenting new results in active areas of formal methods applied to distributed systems.The Programme Committee consisted of M. Broy (Munich, D)B. Moller (Augsburg, D)R. Diaconescu (Bucharest, RO)F. Moller (Uppsala, S)J. Esparza (Munich, D)A. Ponse (Amsterdam, NL)J.F. Groote (Amsterdam, NL)T. Rus (Iowa City, USA)R. Grosu (Philadelphia, USA)E. Stark (Stony Brock, USA)M. Kanovitch (Moscow, RU)Gh. Stefanescu (Bucharest, RO) - chairD. Lucanu (Iasi, RO)D. Walker (Oxford, UK)R. De Nicola (Firenze, I)The Invited Speakers were: Joseph Goguen(University of California at San Diego, USA)Matthew Hennessy (Sussex University, UK)Ugo Montanari (University of Pisa, Italy)The workshop Programme has consisted of three invited lectures, six accepted submissions, and three communications. Except for the papers enclosed in this volume, the program also has included the invited lecture of M. Hennessy, Secure Partial Typing of Mobile Agents in Open Systems and the communications: J.F. Groote and J. vanWamel, Basic theorems for parallel processes in timed miu-CRL; F. Martinelli, A uniform approach for the solution of module checking problems; and C. Zlatea and T. Elrad DUALITY: A Design Model for Mobile Distributed Systems Based on Partial Order Semantics.Organization: The organization of the workshop was partially supported by: Institute of Informatics, Technical University Munich; Department of Fundamentals of Informatics, University of Bucharest; and Faculty of Informatics, Al.I. Cuza University, Iasi. Special thanks go to Prof. Dr. Manfred Broy Chair (Programming Methods, Distributed Systems) at TUM which has supported the cost for the hard copies of this volume.

Assumption-based reasoning in dynamic normative agent systems

In this paper we address dynamic assumption-based reasoning in open agent systems, where, unavoidably, agents have incomplete knowledge about their environment and about other agents. The interactions among agents in such systems are typically subject to norms, which stipulate what each agent is obliged, permitted, prohibited, empowered etc. to do, while it participates in the system. In such environments agents need to resort to assumptions, in order to establish what actions are appropriate to perform, and they need to do so dynamically, since the environment, the agents that exist in it, the information that is exchanged between them, and the normative relations between them change over time. In earlier work, we had proposed Default Theory construction to support dynamic assumption-based reasoning. We argued that in this way, agents could perform both assumption identification and employment dynamically, contrary to other approaches to assumption-based reasoning, which catered for either one or the other. A shortcoming of this early proposal of ours, though, is that Default Theory construction seems to require proof, which is notably computationally expensive. In this paper we present a computational technique that can be used for this construction in an incremental manner that does not depend on proof, and a prototype tool that we developed for experimentation. In a nutshell, depending on their current knowledge at any given time, agents can identify appropriate candidate assumptions in an ad hoc manner. When such choices need to be revised, agents can reconstruct their view of the possible world in which they find themselves, and establish their revised assumption requirements at run-time. This paper is an extended version of the work presented at the 2008 IEEE/WI/ACM Int. Conf. on Intelligent Agent Technology, 9-12 December, Sydney, Australia.

Applying game theory mechanisms in open agent systems with complete information

Game theory is a popular tool for designing interaction protocols for agent systems. It is currently not clear how to extend this to open agent systems. By "open" we mean that foreign agents will be free to enter and leave different systems at will. This means that agents will need to be able to work with previously unseen protocols. There does not yet exist any agreement on a standard way in which such protocols can be specified and published. Furthermore, it is not clear how an agent could be given the ability to use an arbitrary published protocol; the agent would need to be able to work out a strategy for participation. To address this we propose a machine readable language in which a game theory mechanism can be written in the form of an agent interaction protocol. This language allows the workings of the protocol to be made public so that agents can inspect it to determine its properties and hence their best strategy. Enabling agents to automatically determine the game theoretic properties of an arbitrary interaction protocol is difficult. Rather than requiring agents to find the equilibrium of a game, we propose that a recommended equilibrium will be published along with the protocol; agents can then check the recommendation to decide if it is indeed an equilibrium. We present an algorithm for this decision problem. We also develop an equilibrium which simplifies the complexity of the checking problem, while still being robust to unilateral deviations.

Preface: Volume 28

This volume contains the proceedings of the FCT′99 Workshop on Distributed Systems held on 2-3 September 1999 in Iassy, Romania as a satellite event to FCT′99 — the 12th International Symposium on Fundamentals of Computation Theory.Topics: The declared aim of the workshop was to select papers presenting new results in active areas of formal methods applied to distributed systems.The Programme Committee consisted ofM. Broy (Munich, D)B. Moller (Augsburg, D)R. Diaconescu (Bucharest, RO)F. Moller (Uppsala, S)J. Esparza (Munich, D)A. Ponse (Amsterdam, NL)J.F. Groote (Amsterdam, NL)T. Rus (Iowa City, USA)R. Grosu (Philadelphia, USA)E. Stark (Stony Brock, USA)M. Kanovitch (Moscow, RU)Gh. Stefanescu (Bucharest, RO) - chairD. Lucanu (Iasi, RO)D. Walker (Oxford, UK)R. De Nicola (Firenze, I)The Invited Speakers were:Joseph Goguen(University of California at San Diego, USA)Matthew Hennessy (Sussex University, UK)Ugo Montanari (University of Pisa, Italy)The workshop Programme has consisted of three invited lectures, six accepted submissions, and three communications. Except for the papers enclosed in this volume, the program also has included the invited lecture of M. Hennessy, Secure Partial Typing of Mobile Agents in Open Systems and the communications: J.F. Groote and J. vanWamel, Basic theorems for parallel processes in timed miu-CRL; F. Martinelli, A uniform approach for the solution of module checking problems; and C. Zlatea and T. Elrad DUALITY: A Design Model for Mobile Distributed Systems Based on Partial Order Semantics.Organization: The organization of the workshop was partially supported by: Institute of Informatics, Technical University Munich; Department of Fundamentals of Informatics, University of Bucharest; and Faculty of Informatics, Al.I. Cuza University, Iasi. Special thanks go to Prof. Dr. Manfred Broy Chair (Programming Methods, Distributed Systems) at TUM which has supported the cost for the hard copies of this volume.

Behaviour specification of active objects in open generative communication environments

Coordination models based on generative communication are well established for modelling interactions between agents in open systems. Because such models focus on shared data spaces and not on the active agents operating on them, they lack adequate modelling power for specifying agent behaviour. Our work is based on the Objective Linda coordination model (Kielman, 1996) which exploits generative communication and object-based modelling in order to meet the requirements of open systems. For Objective Linda, we present a high-level Petri net formalism for agent behaviour specification that combines an easily understandable, visual representation with the applicability of formal reasoning on agent behaviour. Based on this formalism, we introduce a notion of types and subtyping for active agents. We illustrate the benefits of our work on an example from open systems design.

Developing Real Applications With Agent Technologies

In this paper, we present an agent library to develop multi-agent applications. Such an agent library has been realised taking into account FIPA specifications. FIPA specifications give the minimum amount of technology deemed necessary for the management of agents in an open agent system, including agent roles, an agent communication language, an agent management content language, and a standard way to interact with non-agentised software. Therefore, the library offers the prototypes of the agents necessary for the management of a FIPA agent open system and an agent prototype that the user can specialise to build the other agents necessary for her/his application. The library has been implemented both in C++ and Java. The two implementations support a logical distribution, i.e., the agents correspond to different threads, and a physical distribution, i. e., agents of the same applications run on different machines. The Java library has been used to develop a multi-agent system, that facilitates users to use a video on demand service, and a first prototype of the system is under experimentation. The C++ library is used for the development of a multi-agent system that should integrate the different software modules performing sensory data interpretation, planning, faults diagnosis, . . . of a robot working in a space station.