Showing papers on "Multi-agent system published in 1994"

Interaction and intelligent behavior

Abstract: This thesis addresses situated, embodied agents interacting in complex domains. It focuses on two problems: (1) synthesis and analysis of intelligent group behavior, and (2) learning in complex group environments. Behaviors are proposed as the appropriate level for control and learning. Basic behaviors are introduced as building blocks for synthesizing and analyzing system behavior. The thesis describes the process of selecting such basic behaviors, formally specifying them, algorithmically implementing them, and empirically evaluating them. All of the proposed ideas are validated with a group of up to 20 mobile robots using a basic behavior set consisting of: avoidance, following, aggregation, dispersion, and homing. The set of basic behaviors acts as a substrate for achieving more complex high-level goals and tasks. Two behavior combination operators are introduced, and verified by combining subsets of the above basic behavior set to implement collective flocking and foraging. A methodology is introduced for automatically constructing higher-level behaviors by learning to select among the basic behavior set. A novel formulation of reinforcement learning is proposed that makes behavior selection learnable in noisy, uncertain multi-agent environments with stochastic dynamics. It consists of using conditions and behaviors for more robust control and minimized state-spaces, and a reinforcement shaping methodology that enables principled embedding of domain knowledge with two types of shaping functions: heterogeneous reward functions and progress estimators. The methodology outperforms two alternatives when tested on a collection of robots learning to forage. The proposed formulation enables and accelerates learning in complex multi-robot domains. The generality of the approach makes it compatible with the existing reinforcement learning algorithms, allowing it to accelerate learning in a variety of domains and applications. The presented methodologies and results are aimed at extending our understanding of synthesis, analysis, and learning of group behavior. (Copies available exclusively from MIT Libraries, Rm. 14-0551, Cambridge, MA 02139-4307. Ph. 617-253-5668; Fax 617-253-1690.)

Intelligent Agents: Theory and Practice

Abstract: The concept of an agent has become important in both Artificial Intelligence (AI) and mainstream computer science. Our aim in this paper is to point the reader at what we perceive to be the most important theoretical and practical issues associated with the design and construction of intelligent agents. For convenience, we divide these issues into three areas (though as the reader will see, the divisions are at times somewhat arbitrary). Agent theory is concerned with the question of what an agent is, and the use of mathematical formalisms for representing and reasoning about the properties of agents. Agent architectures can be thought of as software engineering models of agents;researchers in this area are primarily concerned with the problem of designing software or hardware systems that will satisfy the properties specified by agent theorists. Finally, agent languages are software systems for programming and experimenting with agents; these languages may embody principles proposed by theorists. The paper is not intended to serve as a tutorial introduction to all the issues mentioned; we hope instead simply to identify the most important issues, and point to work that elaborates on them. The article includes a short review of current and potential applications of agent technology.

Cooperation in industrial multi-agent systems

Abstract: Initial experiences of building industrial multi-agent systems representing co-ordination in multi-agent systems Grate*: a co-operation knowledge level multi-agent system conclusion and future directions.

Distributed Problem Solving and Multi-Agent Systems: Comparisons and Examples*

Abstract: The term multi-agent system is currently in vogue, and has been generally applied to any system that is, or can be considered to be, composed of multiple interacting agents. In the various multi-agent (or, more properly, multiagent) systems that have been proposed developed, a wide variety of "agents" have been considered, ranging from fully autonomous intelligent agents (such as people) down to relatively simple entities (such as rules or clusters of rules). Thus, as it has become progressively used, "multiagent systems" has come to encompass an increasingly broad variety of issues, approaches, and phenomena, to the point

An architecture for dynamically interacting agents

Abstract: In this paper, we present the layered agent architecture INTERRAP which has been developed to cope with the basic requirements for Cooperative Intelligent Systems (CIS): agents shall behave in a situated, efficient, and goal-directed manner, and they shall be able to interact (i.e. coordinate and collaborate) with other agents. Over the past few years, agent architectures combining reactive with deliberative facilities have become very trendy in DAI. However, most attempts end up at the local planning layer and fail to take into account important mechanisms for interaction and collaboration among agents. INTERRAP extends previous attempts to build layered architectures by a cooperation component which holds cooperation knowledge as well as a library of joint plant. The modules of the model and the flow of control among them are explained. The model is described and evaluated by the FORKS application, the simulation of an automated loading-dock.

ARCHON: framework for intelligent cooperation

Abstract: The paper describes the work that has taken place in the ARCHON Project, ESPRIT project P-2256. The consortium has developed a general-purpose architecture, which can be used to facilitate cooperative problem-solving in industrial applications. The paper describes the need for a multiple agent approach for industrial applications, outlines the benefits which can be accrued by adopting this paradigm, and describes the key difficulties which must be faced when building a multi-agent system in this domain. Details of the ARCHON architecture are presented, including a description of the main functional components and their realisation in a hybrid agent model. An example of cooperative fault diagnosis in an electricity management application is described in order to provide a clear illustration of the working of the ARCHON architecture, and to provide a concrete example of the potential benefits of a multi-agent approach. >

Belief revision in multi-agent systems

Abstract: The ability to respond sensibly to changing and conflicting beliefs is an integral part of intelligent agency. To this end, we outline the design and implementation of a Distributed Assumption-based Truth Maintenance Sys- tem (DATMS) appropriate for controlling cooperative problem solving in a dynamic real world multi-agent community. Our DATMS works on the prin- ciple of local coherence which means that different agents can have different perspectives on the same fact provided that these stances are appropriately justified. The belief revision algorithm is presented, the meta-level code needed to ensure that all system-wide queries can be uniquely answered is described, and the DATMS' implementation in a general purpose multi-agent shell is discussed.

Organizational intelligence and negotiation based DAI systems: Theoretical foundations and experimental results

Abstract: A steadily increasing number of researchers believes that so-called organizational multi agent systems are a key technology to support information and knowledge processing activities in cooperative, networked organizations. This, in turn, necessitates their integration with the underlying human-centred organization. The concept of an organization has emerged as central to the structuring of activities of both decentralized industrial and commercial conglomerates and collections of intelligent problem solvers within Distributed Artificial Intelligence (DAI) systems. Of late a new discipline has begun to emerge, that of Organizational Intelligence (OI). Organizational Intelligence demands a greater synthesis between the principles of Organization Theory (OT) and DAI, by the explicit incorporation of theories of both organizations and DAI into the field of OI. This paper concentrates on two rather important features of OI, namely organizational memory and learning capabilities. It will first discuss the theoretical foundations. Then it will be shown how the contract net approach can be extended to meet these demands. Finally, it will be proved by some experimental results that the increased intellectual capabilities of the extended contract net will substantially contribute to the performance as well as the quality of solution processes.

An agent model combining reactive and cognitive capabilities

Abstract: Presents an abstract model of an autonomous agent that integrates reactive and cognitive behaviour. The authors state three requirements (reactivity, timely behaviour, and symbolic representation) that should be accomplished by such an integration. Due to the fact that there exists no model up to now that covers all three aspects, the authors present an abstract model of an autonomous agent that does meet these requirements. To the modules of knowledge representation, planning, scheduling, and execution, the authors add the process of evaluation that constantly supervises the environment and the agent's actions in order to ensure the agent's reactivity. To validate the proposed model, the authors discuss an implementation in a complex robot scenario. >

A model for intent interpretation for multiple agents with conflicts

Abstract: Intent interpretation is the process of forming a causal explanation of the observed actions of an agent in terms of the current purposes of the agent. Most of the current work in this area has dealt with interpretation of the actions of single agents without unresolved conflicts. This paper describes the use of the plan-goal representational approach as originally formulated in Geddes (1989) to interpret actions of multiple agents with unresolved conflicts. This situation is of importance in the practical world, arising often in cooperative team and organizational behavior as well as in overt adversarial behavior. >

MAVI: a multi-agent system for visual integration

Abstract: This paper presents a control architecture for building integrated vision systems. This system is based on a multi-agent approach which allows the definition of an open and flexible architecture for the integration of visual modules. It is shown how this architecture permits the dynamic definition of the control with some visual goals to satisfy in the context of an active vision system. >

Plan execution motivation in multi-agent systems

Abstract: In this paper we analyze a particular model of control among intelligent agents, that of non-absolute control. Non-absolute control involves a "supervisor" agent that issues orders to a group of "subordinate" agents. An example might be an Internet user who issues a query to a group of software agents on remote hosts, or a human agent on Earth directing the activities of Mars-based semi-autonomous vehicles. The members of the subordinate group are assumed to be self-motivated, and individually rational (i.e., they are basically willing to carry out the supervisor's request if properly compensated). This assumption gives rise to the need for a reward policy that would motivate each agent to contribute to the group activity. In this paper we introduce such a policy under certain simplifying assumptions.

Use of multi agent systems for mobile robotics control

Abstract: Multi-agent systems in mobile robotics are generally use to represent multirobot systems and to study the interaction between robots to achieve a global task. But mobile robot agents are designed in different ways, with a reactive level and sometimes a cognitive level to provide the behavior of the agent with a deliberative decision making. Alter a brief survey of the use of multi-agent systems in mobile robotics, we explore the possibility to design a mobile robot agent using another multi-agent system based on self-interested autonomous entities of the mobile robot. >

Cooperative multi-agent system for an assembly robotics cell

Abstract: A multi-agent system architecture is described and justified for the sake of its application to an assembly robotics testbed. A blackboard-based agent using a GoalBlackboard/DataBlackboard facility for intra (not inter) agent communication and including knowledge about all the agent's community, is presented as well as its main functionality. Coordination of different agents dynamically playing the roles either of organizers or respondents may lead to the use of either negotiation or client/server protocols for cooperation. Also come cooperative strategies and involved knowledge have been studied, classified and implemented in the robotics testbed enabling a sophisticated agent behavior both in terms of cooperation and local control. A real testbed, whose agents are briefly presented here, working with real-time constraints, has already been implemented and tested in our laboratory.

DYNACLIPS (DYNAmic CLIPS): A dynamic knowledge exchange tool for intelligent agents

Abstract: In a dynamic environment, intelligent agents must be responsive to unanticipated conditions. When such conditions occur, an intelligent agent may have to stop a previously planned and scheduled course of actions and replan, reschedule, start new activities and initiate a new problem solving process to successfully respond to the new conditions. Problems occur when an intelligent agent does not have enough knowledge to properly respond to the new situation. DYNACLIPS is an implementation of a framework for dynamic knowledge exchange among intelligent agents. Each intelligent agent is a CLIPS shell and runs a separate process under SunOS operating system. Intelligent agents can exchange facts, rules, and CLIPS commands at run time. Knowledge exchange among intelligent agents at run times does not effect execution of either sender and receiver intelligent agent. Intelligent agents can keep the knowledge temporarily or permanently. In other words, knowledge exchange among intelligent agents would allow for a form of learning to be accomplished.

An Agent-based Architecture for Concurrent Engineering

Abstract: This paper describes a reference architecture for a concurrent engineering environment. The architecture embeds applications in wrappers to treat the applications as individual reasoning agents. Embedded applications have included planning, vision, simulation, and robotic control modules. These agents negotiate from various perspectives to arrive at globally acceptable solutions. The agents transmit information and negotiate using a canonical vocabulary based upon an international standard for data exchange. Because our implementation of the architecture explicitly considers real-world issues, such as heterogeneous systems, distributed processing, faults, meeting real-time deadlines, and communication delays, the implementation of architecture facilitates rapid transferral from laboratory experiments to field systems.

Hierarchical Model and Communication by Signs, Signals and Symbols in Multi-Agent Environments

Abstract: This paper describes a general framework for designing agents for a multiagent systems. A hierarchical agent model is described, structuring an agent according to different types of situations. It has to deal with: routines, familiar and unfamiliar situations. Then, an idea is developped for the coordination between agents: agents should prefer low levels (i.e. routine and familiar situations) than the high level (i.e. unfamiliar situations). The reason is that low levels are fast, effortless and are propitious for coordinated activities between agents, whereas the high level is slow, laborious and can lead to conflicts between agents. To achieve this, we develop the idea of using communication by signs and signals (and not by symbols) in order to allow agents to rely on their low levels. Finally, implementation and experiments demonstrated, on some scenarios of urban traffic, the applicability of concepts developed in this article.

Group behavior and group learning

Abstract: We describe an approach to principled synthesis and analysis of group behavior in situated, embodied multiagent systems. We propose basic behaviors as the appropriate level for control and learning. Basic behaviors are generated by simple local rules and serve as building blocks for a large repertoire of higher level behaviors. We describe an architecture for combining basic behaviors into compound, more complex tasks. We also describe a formulation of reinforcement learning that allows for learning such compound behaviors automatically in non Markovian, noisy and uncertain environments with multiple agent. We demonstrate all of our methodologies with experimental data on a collection of physical mobile robots demonstrating group avoidance, aggregation, dispersion, following, wandering, flocking, and foraging.

Agentifying the process: task-based or robot-based decomposition?

Abstract: In this paper we advocate the use of multi-agent systems in cellular manufacturing. More specifically, we argue that multi-agent systems can respond to the flexibility and adaptivity needs of modern manufacturing. The central issue of the design of such a multi-agent manufacturing system is the choice of the granularity level: are agents defined by decomposing the whole process across tasks or across robots? We argue that at different levels of the production plant we need different decomposition grains, robot-based, task-based or hybrid, and that this granularity decision should be driven by the criteria of control localisation, knowledge decoupling and interaction minimisation, so as to identify the decision points of the overall process. Those criteria define a dual composition and decomposition relation that appears as a good substrate for system scaling. >

MAPS: a language for multi-agent system design

Abstract: A language for multi-agent system design (MAPS) is presented and discussed in this paper. Any agent in MAPS is thought of as an expert system standing on its own. It can communicate through synchronous and asynchronous message sending. Dedicated behaviours are provided which specify how incoming messages are processed. Inter-agent cooperation is controlled via production rules. Two pre-defined agent classes are provided, which are given specific problem-solving roles: Knowledge Server (KS) agents are meant to maintain and transmit knowledge about problem-solving states, while Knowledge Processor (KP) agents are meant to process these elements in order to progress towards a solution. The duality between agent and resource modelling levels on the one hand and between KS and KP modelling styles on the other is shown to allow the specification of various control strategies. The environment is currently running on HP, SUN and DEC workstations.

A Tool for Handling Uncertain Information in Multi-Agent Systems

Abstract: We focus on the strategy problem arising as a manager faces situations involving a choice between a finite set of strategies, having access to a finite set of agents reporting their opinions. The most preferred strategy is determined from the agents' individual opinions and the relative credibility of each agent. The evaluation method used is primarily based on the principle of maximising the expected utility. The evaluation results in a set of admissible strategies. These strategies can be further investigated with respect to their relative strengths and also with respect to the number of values consistent with the given domain that make them admissible.

Combining predictive and reactive control strategies in multi-agent systems

Abstract: Traditionally, researchers into robot control have applied one of two broad control strategies; predictive control or reactive control. This paper proposes a new control strategy for multi-robot systems which combines the predictive and reactive styles with the aim of drawing on the strengths of both; the strong task definition capability of the predictive approach with the robustness and reliability of the reactive approach. A reactive task execution architecture, the Behaviour Synthesis Architecture, has been fully developed and implemented on two mobile robots, which co-operatively execute local tasks, including the joint relocation of pallets within a semi-structured environment. The remainder of the proposed architecture is under development as part of an ongoing research programme.< >

Towards a Dynamic Multi-Agent Organization

Abstract: A method enabling the dynamic organization of multiagent systems is proposed. It is based on studies about performance of different organizations when they have to handle various situations. The method uses a library of organizational models and three selection criteria relative to the current task the multiagent has to achieve. The agents apply the selected organizational model by the propagation of behavioral rules among them. Endly, some implementation aspects are presented.

Multi-agent system

Abstract: PURPOSE: To provide the system which is highly flexible and has self-organizing capability by automatically adding a new interposing agent. CONSTITUTION: An agent information generating means 5 which generates new agent information by utilizing an evaluation signal is added to the multi- agent system equipped with an input agent 1, an output agent 3, and interposing agents 2, and an additional agent consisting of an input/output buffer 13 which exchanges a communication signal with other agents, an agent information storage part 12 which holds the agent information, a data storage part 11 which stores data with evaluation as a set of the state of the input buffer a and the evaluation signal, and an input/output device 19 which updates the contents of the input/output buffer with the data with evaluation by utilizing the agent information is added sequentially as an interposing agent.

A FRAMEWORK FOR DYNAMIC KNOWLEDGE EXCHANGE AMONG INTELLIGENT AGENTS Authors

Abstract: An intelligent agent is an object with its own knowledge and information base. F, ach intelligent agent acts in parallel with other intelligent agents and cooperates with a selected set of other agents to achieve a common set of goals. In a dynamic environment, intelligent agents must be responsive to unanticipated conditions. When such conditions occur, an agent may have to stop a previously planned and scheduled course of actions and replan, reschedulc, start new activities and initiate a new problem solving process to successfully respond to the new conditions. Problems occur when an intelligent agent does not have enough knowledge to properly reslxmd to the new situation. A framework lbr dynamic knowledge exchange among intelligent agents has been proposed to allow an intelligent agent to react to knowledge deficiency. Hence, using the prolx)sed framework new knowledge can be transferred when an intelligent agent is unable to solve the problem using its own knowledge. Once the knowledge has been transferred, the intelligent agent can either keep the transferred knowledge permanently or remove it after the transferred knowledge has he, en used to re)lye the problem

Extending a parallel CLP language to support the development of multi-agent systems

Abstract: An extension of the parallel constraint logic programming language ElipSys is presented. This extension is directed towards the development of multi-agent systems which have to deal with large combinatorial problems that are distributed in nature. Problems of this kind, after being decomposed into subproblems, may be tackled efficiently by individual agents using ElipSys’ powerful mechanisms, such as parallelism and constraint satisfaction techniques. The proposed extension supports the communication requirements of the agents, in order to have them cooperate and solve the original combinatorially intensive problem. The communication scheme among the agents is viewed as a three-layered model. The first layer is socket oriented, the second realizes a blackboard architecture and the third supports virtual point-topoint interaction among the agents.

Procedure-oriented studies of collective phenomena in multi-agent systems

Abstract: The research described in this paper concerns alliance formation and norms of behavior in communities of intelligent agents. An alliance is composed of intelligent agents that have made a deliberate decision to join forces to achieve alliance goals. Norms are defined as limitations on agents, such as legal constraints on individual actions. The aim of this research is to arrive at descriptive and prescriptive theories of group behavior that explain these phenomena. This paper describes information structures and procedural methods that allow the development of such a procedural model. These studies will contribute to the understanding of human organizations, including political activity, national and international markets, and other social contexts. >