An Automated Teamwork Infrastructure for Heterogeneous Software Agents and Humans

Agent integration architectures enable a heterogeneous, distributed set of agents to work together to address problems of greater complexity than those addressed by the individual agents themselves. Unfortunately, integrating software agents and humans to perform real-world tasks in a large-scale system remains difficult, especially due to three main challenges: ensuring robust execution in the face of a dynamic environment, providing abstract task specifications without all the low-level coordination details, and finding appropriate agents for inclusion in the overall system. To address these challenges, our Teamcore project provides the integration architecture with general-purpose teamwork coordination capabilities. We make each agent team-ready by providing it with a proxy capable of general teamwork reasoning. Thus, a key novelty and strength of our framework is that powerful teamwork capabilities are built into its foundations by providing the proxies themselves with a teamwork model.Given this teamwork model, the Teamcore proxies addresses the first agent integration challenge, robust execution, by automatically generating the required coordination actions for the agents they represent. We can also exploit the proxies' reusable general teamwork knowledge to address the second agent integration challenge. Through team-oriented programming, a developer specifies a hierarchical organization and its goals and plans, abstracting away from coordination details. Finally, KARMA, our Knowledgeable Agent Resources Manager Assistant, can aid the developer in conquering the third agent integration challenge by locating agents that match the specified organization's requirements. Our integration architecture enables teamwork among agents with no coordination capabilities, and it establishes and automates consistent teamwork among agents with some coordination capabilities. Thus, team-oriented programming provides a level of abstraction that can be used on top of previous approaches to agent-oriented programming. We illustrate how the Teamcore architecture successfully addressed the challenges of agent integration in two application domains: simulated rehearsal of a military evacuation mission and facilitation of human collaboration.     

Automated Testing of Physical Security: Red Teaming Through Machine Learning

Modern surveillance systems for practical applications with diverse and mobile sensors are large, complex, and expensive. It is known that unexpected behaviors can emerge from such systems, and when these behaviors correspond to weaknesses in a surveillance system, we call them emergent vulnerabilities. Given their cost and importance to security, it is essential to test these systems for such vulnerabilities prior to deployment. To that end, we automate the testing process with multiagent systems and machine learning. However, the conventional—and most intuitive–approach is to focus the machine learning on the subject system, which leads to a high‐dimensional problem that is intractable. Instead, we demonstrate in this paper that learning attacks on the system is tractable and provides a viable testing method. We demonstrate this with a series of studies in simulation and with a small‐scale model system featuring elements typically found in real physical surveillance systems. Our machine learning method finds successful attacks in simulation, which we can duplicate with the physical system. The method is scalable, with the implication that it could be used to test larger, real surveillance installations.     

An ontology for commitments in multiagent systems:

Social commitments have long been recognized as an important concept for multiagent systems. We propose a rich formulation of social commitments that motivates an architecture for multiagent systems, which we dub spheres of commitment. We identify the key operations on commitments and multiagent systems. We distinguish between explicit and implicit commitments. Multiagent systems, viewed as spheres of commitment (SoComs), provide the context for the different operations on commitments. Armed with the above ideas, we can capture normative concepts such as obligations, taboos, conventions, and pledges as different kinds of commitments. In this manner, we synthesize ideas from multiagent systems, particularly the idea of social context, with ideas from ethics and legal reasoning, specifically that of directed obligations in the Hohfeldian tradition.     

Agents in Electronic Commerce: Component Technologies for Automated Negotiation and Coalition Formation

Automated negotiation and coalition formation among self-interested agents are playing an increasingly important role in electronic commerce. Such agents cannot be coordinated by externally imposing their strategies. Instead the interaction protocols have to be designed so that each agent is motivated to follow the strategy that the protocol designer wants it to follow. This paper reviews six component technologies that we have developed for making such interactions less manipulable and more efficient in terms of the computational processes and the outcomes: 1. OCSM-contracts in marginal cost based contracting, 2. leveled commitment contracts, 3. anytime coalition structure generation with worst case guarantees, 4. trading off computation cost against optimization quality within each coalition, 5. distributing search among insincere agents, and 6. unenforced contract execution. Each of these technologies represents a different way of battling self-interest and combinatorial complexity simultaneously. This is a key battle when multi-agent systems move into large-scale open settings.     

A group‐oriented secure multiagent platform

Security is becoming a major concern in multiagent systems, since an agent's incorrect or inappropriate behaviour may cause non‐desired effects, such as money and data loss. Some multiagent platforms (MAP) are now providing baseline security features, such as authentication, authorization, integrity and confidentiality. However, they fail to support other features related to the sociability skills of agents such as agent groups. What is more, none of the listed MAPs provide a mechanism for preserving the privacy of the users (regarding their identities) that run their agents on such MAPs. In this paper, we present the security infrastructure (SI) of the Magentix MAP, which supports agent groups and preserves user identity privacy. The SI is based on identities that are assigned to all the different entities found in Magentix (users, agents and agent groups). We also provide an evaluation of the SI describing an example application built on top of Magentix and a performance evaluation of it. Copyright © 2010 John Wiley & Sons, Ltd.     

Science of Scientific Team Science: A survey

Scientific teamwork collaboration is an integral element of the scientific process that often leads to significant findings. Systematic analysis of scientific teamwork collaboration continues to influence both the advance in science and knowledge production. This paper presents an overview of Science of Scientific Team Science (SSTS). SSTS explores the behaviors and attributes of teamwork and team-based collaboration specific to scientific teams from the perspective of quantitative analysis, which refers to a branch of science that analyzes and discovers scientific collaboration patterns inter- or extra-team. Aiming at assisting scientific team formation, improving collaboration environment, evaluating team performance, and fostering collaborative behaviors, this survey presents an overview in SSTS. Theoretical background of SSTS at different team development stages has been discussed. In addition, three classifications of SSTS, including interdisciplinary, multidisciplinary, and transdisciplinary research approaches have been investigated. Their associated similarities and differences, challenges and benefits, are also examined. This paper also summarizes web-based tools that enhance one’s understanding and opinion of SSTS. Key technologies and open issues are then discussed. The association among scientific collaboration, scientific teamwork, SSTS, and cross-disciplinary research gives rise to critical implications for scholars who wish to employ and invest in those issues.     

Real-Time Search for Autonomous Agents and Multiagent Systems

Since real-time search provides an attractive framework for resource-bounded problem solving, this paper extends the framework for autonomous agents and for a multiagent world. To adaptively control search processes, we propose -search which allows suboptimal solutions with error, and -search which balances the tradeoff between exploration and exploitation. We then consider search in uncertain situations, where the goal may change during the course of the search, and propose a moving target search (MTS) algorithm. We also investigate real-time bidirectional search (RTBS) algorithms, where two problem solvers cooperatively achieve a shared goal. Finally, we introduce a new problem solving paradigm, called organizational problem solving, for multiagent systems.     

A general framework for coherence in a CKBS

Coherence in a distributed system is meant to offset the disadvantages of distribution. The paper explores four issues under coherence, namely preservation of knowledge consistency across the agents, reliability of the overall system, integration of local solutions and the global performance. It presents some general strategies that can be employed to improve coherence in a CKBS, which include a weak consistency with versions for knowledge revision, and a recovery mechanism based on a hierarchic three-stage coordination, which ensures the correct isolation of potentially hierarchic multiagent actions. The paper goes on to identify the sources and classes of conflicts in global integration, and it suggests remedies, which at worst case would involve negotiation. In global performance, it focusses on planning and result synthesis, as the two most important problem domains, and suggests strategies ameliorate performance.     

一个基于XCS的同质团队学习模型

同质团队学习是实现多Agent协作的一种方法。但是,传统方法仅在系统运行的前后对目标Agent进行修改,系统运行过程没有直接对Agent的改进做出贡献。本文利用合作策略,在学习分类器系统XCS的基础上提出了一种同质团队学习模型,弥补了传统方法的上述不足。文中还在模型的基础上实验分析了相关因素,如规则积累、通信以及发现新规则等对多Agent协作效率的影响。     

Applying Gaia and AUML for the development of multiagent‐based control software for flexible manufacturing systems: addressing methodological and implementation issues

In this article, we present the development of a simple multiagent‐based system for the control of a flexible manufacturing system. We followed the stages of a methodology specially conceived for the development of agent‐based system, which is an integration of the classical methodology for agent‐oriented analysis and design Gaia, and AUML (Agent‐Unified Modeling Language). We adopted as study case the CIMUBB Laboratory at the University of Bio‐Bio, which has a flexible manufacturing system including three flexible manufacturing cells interconnected by a conveyor belt. In the analysis stage, we identified roles involved, and we design models representing roles and protocols. In the design stage, we applied Gaia agent, services, and acquaintance models from Gaia, and we complemented with AUML as the adopted methodology suggests. With the developed models, we constructed a fully functional system where each agent was built as an independent process tree. Agents communicate by passing messages through the Ethernet network with socket interfaces. Various tests executed in our laboratory scale manufacturing system show the effectiveness of our implementation. Copyright © 2015 John Wiley & Sons, Ltd.     

Level-of-detail for cognitive real-time characters

We present a solution for the real-time simulation of artificial environments containing cognitive and hierarchically organized agents at constant rendering framerates. We introduce a level-of-detail concept to behavioral modeling, where agents populating the world can be both reactive and proactive. The disposable time per rendered frame for behavioral simulation is variable and determines the complexity of the presented behavior. A special scheduling algorithm distributes this time to the agents depending on their level-of-detail such that visible and nearby agents get more time than invisible or distant agents. This allows for smooth transitions between reactive and proactive behavior. The time available per agent influences the proactive behavior, which becomes more sophisticated because it can spend time anticipating future situations. Additionally, we exploit the use of hierarchies within groups of agents that allow for different levels of control. We show that our approach is well-suited for simulating environments with up to several hundred agents with reasonable response times and the behavior adapts to the current viewpoint.     

An Infrastructure for Argumentative Agents

Multiagent systems are suitable for providing a framework that allows agents to perform collaborative processes in a social context. Furthermore, argumentation is a natural way of reaching agreements between several parties. However, it is difficult to find infrastructures of argumentation offering support for agent societies and their social context. Offering support for agent societies allows representation of more realistic environments to have argumentation dialogues. We propose an infrastructure to develop and execute argumentative agents in an open multiagent system. It offers tools to develop agents with argumentation capabilities. It also offers support for agent societies and their social context. The infrastructure is publicly available. Also, it has been implemented in an application scenario where argumentative agents try to reach an agreement about the best solution to solve a problem reported to the system.     

Nussbaum function–based universal cooperative output regulation design for uncertain nonlinear multiagent systems

This paper presents a Nussbaum function–based universal cooperative output regulation design for a class of nonlinear multiagent systems with both an unknown exosystem and nonidentical unknown control directions. The major challenges include the nonidentical unknown control directions in a directed communication graph and the concurrence of the unknown parameters in both the plant and the exosystem. To handle the nonidentical unknown control directions, we propose a dynamic compensator–based distributed controller such that the Nussbaum gain technique can be successfully implemented under directed communication graphs. Moreover, to deal with the unknown exosystem, we integrate the distributed controller with a novel internal model candidate. The resulting distributed controller is a universal regulator in the sense that it does not require the unknown parameters to be in known compact sets. Furthermore, the proposed controller is more flexible compared with those in the existing works as any existing Nussbaum gains can be adopted in the controller design and the adopted Nussbaum gains can be nonidentical for each agent.     

Allocating structured tasks in heterogeneous agent teams

Task allocation is an important aspect of multiagent coordination. However, there are many challenges in developing appropriate strategies for multiagent teams so that they operate efficiently. Real‐world scenarios such as flooding disasters usually require the use of heterogeneous robots and the execution of tasks with different structures and complexities. In this paper, we propose a decentralized task allocation mechanism considering different types of tasks for heterogeneous agent teams where agents play different roles and carry out tasks according to their own capabilities. We have run several experiments to evaluate the proposed mechanism. The results show that the proposed mechanism appears to scale well and provides near‐optimal allocations.     

Stability analysis of nonholonomic multiagent coordinate‐free formation control subject to communication delays

This paper investigates the convergence of nonholonomic multiagent coordinate‐free formation control to a prescribed target formation subject to communication delays by means of Lyapunov‐Krasovskii approach and smooth state‐feedback control laws. As a result, an iterative algorithm based on linear matrix inequalities is provided to obtain the worst‐case point‐to‐point delay under which the multiagent system is guaranteed to be stable. It is worth mentioning that: (i) the given algorithm holds for any connected communication topology and (ii) the formation control is coordinate‐free, that is, a common frame is not required to be shared between agents. The effectiveness of the given method is illustrated through simulation results.     

DART: A DISTRIBUTED ANALYSIS OF REPUTATION AND TRUST FRAMEWORK

Artificial societies—distributed systems of autonomous agents—are becoming increasingly important in open distributed environments, especially in e‐commerce. Agents require trust and reputation concepts to identify communities of agents with which to interact reliably. We have noted in real environments that adversaries tend to focus on exploitation of the trust and reputation model. These vulnerabilities reinforce the need for new evaluation criteria for trust and reputation models called exploitation resistance which reflects the ability of a trust model to be unaffected by agents who try to manipulate the trust model. To examine whether a given trust and reputation model is exploitation‐resistant, the researchers require a flexible, easy‐to‐use, and general framework. This framework should provide the facility to specify heterogeneous agents with different trust models and behaviors. This paper introduces a Distributed Analysis of Reputation and Trust (DART) framework. The environment of DART is decentralized and game‐theoretic. Not only is the proposed environment model compatible with the characteristics of open distributed systems, but it also allows agents to have different types of interactions in this environment model. Besides direct, witness, and introduction interactions, agents in our environment model can have a type of interaction called a reporting interaction, which represents a decentralized reporting mechanism in distributed environments. The proposed environment model provides various metrics at both micro and macro levels for analyzing the implemented trust and reputation models. Using DART, researchers have empirically demonstrated the vulnerability of well‐known trust models against both individual and group attacks.     

Learning to Improve Coordinated Actions in Cooperative Distributed Problem-Solving Environments

Coordination is an essential technique in cooperative, distributed multiagent systems. However, sophisticated coordination strategies are not always cost-effective in all problem-solving situations. This paper presents a learning method to identify what information will improve coordination in specific problem-solving situations. Learning is accomplished by recording and analyzing traces of inferences after problem solving. The analysis identifies situations where inappropriate coordination strategies caused redundant activities, or the lack of timely execution of important activities, thus degrading system performance. To remedy this problem, situation-specific control rules are created which acquire additional nonlocal information about activities in the agent networks and then select another plan or another scheduling strategy. Examples from a real distributed problem-solving application involving diagnosis of a local area network are described.     

Complementary‐based coalition formation for energy microgrids

In recent years, the notion of electrical energy microgrids (MGs), in which communities share their locally generated power, has gained increasing interest. Typically, the energy generated comes from renewable resources, which means that its availability is variable, ie, sometimes there may be energy surpluses and at other times energy deficits. This energy variability can be ameliorated by trading energy with a connected electricity grid. However, since main electricity grids are subject to faults or other outages, it can be advantageous for energy MGs to form coalitions and share their energy among themselves. In this work, we present our model for the dynamic formation of such MG coalitions. In our model, MGs form coalitions on the basis of complementary weather patterns. Our agent‐based model, which is scalable and affords autonomy among the MGs participating in the coalition (agents can join and depart from coalitions at any time), features methods to reduce overall “discomfort” so that, even when all participating MGs in a coalition experience deficits, they can share energy so that their overall discomfort is reduced. We demonstrate the efficacy of our model by showing empirical studies conducted with real energy production and consumption data.     

Predicting teamwork results from social network analysis

Modelling students' behaviours has reached a status that can only be overcome by improving the ability of predicting the results on teamwork. Indeed, teamwork is an important piece on the learning process, but understanding their mechanisms and predicting the results achieved is far from being solved by traditional classifiers. In this paper, we address the problem of predicting teamwork results, and propose a recommender system that suggests new teams, in the context of a given curricular unit. Any student, who is looking for a team, may use the system; in particular, he may ask for the best team to join, either considering all available colleagues or just the set of his previous teammates. Our system makes use of social network analysis and classification methods as the algorithmic core of the decision‐making process. System evaluation is presented through a set of experimental results, which report the performance of social network analysis and classification algorithms over real datasets.