Multi-agent Systems Community Research Articles

The Fairness Fair: Bringing Human Perception into Collective Decision-Making

Fairness is one of the most desirable societal principles in collective decision-making. It has been extensively studied in the past decades for its axiomatic properties and has received substantial attention from the multiagent systems community in recent years for its theoretical and computational aspects in algorithmic decision-making. However, these studies are often not sufficiently rich to capture the intricacies of human perception of fairness in the ambivalent nature of the real-world problems. We argue that not only fair solutions should be deemed desirable by social planners (designers), but they should be governed by human and societal cognition, consider perceived outcomes based on human judgement, and be verifiable. We discuss how achieving this goal requires a broad transdisciplinary approach ranging from computing and AI to behavioral economics and human-AI interaction. In doing so, we identify shortcomings and long-term challenges of the current literature of fair division, describe recent efforts in addressing them, and more importantly, highlight a series of open research directions.

Toward cooperative multi-agent video streaming perception

Video streaming perception ability is critical for AI applications on resource-constrained devices (agents), which prefers to offload video streams from devices to edge servers for real-time inference by deep neural networks (DNNs). Meanwhile, the multi-agent system (MAS) community is attempting to run DNNs on multiple cooperative agents to enable improved swarm intelligence-based tasks (e.g., drone swarm intelligence, self-driving fleet collaboration, and multi-agent robot cooperation). However, transferring video streaming perception capability from single-agent systems to MASs is extremely difficult due to spontaneous competition-induced trade-offs between the desired goals of accuracy, consistency, and capacity, which are three critical but conflicting measuring indexes. In this paper, we present the design and implementation of MASSIVE, an edge-assisted cooperative multi-agent video streaming perception system that simultaneously achieves all three desired goals. In our design, we consider the performance characteristics of video streaming perception and the insight of its periodic offloading pattern. On this basis, we develop a Pareto improvement scheduler to eliminate spontaneous competition among agents, allowing multi-objective optimization to achieve an ideal Pareto optimal state. Finally, we propose a virtual traffic shaper based on the mainstream 802.11 MAC protocol to ensure deterministic periodic video stream offloading in an uncertain wireless network. Our experiments demonstrate that MASSIVE achieves 122.7% accuracy and 1.8x capacity compared to the closest baseline on multiple actual cooperative vision tasks with even better consistency, and achieves an ideal Pareto optimal state in a wireless environment.

Enabling imitation-based cooperation in dynamic social networks

The emergence of cooperation among self-interested agents has been a key concern of the multi-agent systems community for decades. With the increased importance of network-mediated interaction, researchers have shifted the attention to the impact of social networks and their dynamics in promoting or hindering cooperation, drawing various context-dependent conclusions. For example, some lines of research, theoretical and experimental, suggest the existence of a threshold effect in the ratio of timescales of network evolution, after which cooperation will emerge, whereas other lines dispute this, suggesting instead a Goldilocks zone. In this paper we provide an evolutionary game theory framework to understand coevolutionary processes from a bottom up perspective - in particular the emergence of a cooperator-core and defector-periphery - clarifying the impact of partner selection and imitation strategies in promoting cooperative behaviour, without assuming underlying communication or reputation mechanisms. In doing so we provide a unifying framework to study imitation-based cooperation in dynamic social networks and show that disputes in the literature can in fact coexist in so far as the results stem from different equally valid assumptions.

Impartial Selection with Additive Approximation Guarantees

Impartial selection has recently received much attention within the multi-agent systems community. The task is, given a directed graph representing nominations to the members of a community by other members, to select the member with the highest number of nominations. This seemingly trivial goal becomes challenging when there is an additional impartiality constraint, requiring that no single member can influence her chance of being selected. Recent progress has identified impartial selection rules with optimal approximation ratios. Moreover, it was noted that worst-case instances are graphs with few vertices. Motivated by this fact, we propose the study of additive approximation, the difference between the highest number of nominations and the number of nominations of the selected member, as an alternative measure of the quality of impartial selection. Our positive results include two randomized impartial selection mechanisms which have additive approximation guarantees of $\Theta(\sqrt{n})$ and $\Theta(n^{2/3}\ln^{1/3}n)$ for the two most studied models in the literature, where $n$ denotes the community size. We complement our positive results by providing negative results for various cases. First, we provide a characterization for the interesting class of strong sample mechanisms, which allows us to obtain lower bounds of $n-2$, and of $\Omega(\sqrt{n})$ for their deterministic and randomized variants respectively. Finally, we present a general lower bound of $2$ for all deterministic impartial mechanisms.

Computing Game-Theoretic Solutions and Applications to Security

The multiagent systems community has adopted game theory as a framework for the design of systems of multiple self-interested agents. For this to be effective, efficient algorithms must be designed to compute the solutions that game theory prescribes. In this paper, I summarize some of the state of the art on this topic, focusing particularly on how this line of work has contributed to several highly visible deployed security applications, developed at the University of Southern California.

Tendencies in Multi-Agent Systems: A Systematic Literature Review

The application of Artificial Intelligence mechanisms allows the development of systems capable to solve very complex engineering problems. Multi-agent systems (MAS) are one paradigm that allows an alternative way to design distributed control systems. While research in this area grew exponentially before 2009, there is a need to understand the status quo of the field from 2009 to June 2017. An extension of the results of a SLR related to Multi-Agent Systems, its applications and research gaps, following Kitchenham and Wholin guidelines are presented in this paper. From the analysis of 279 papers (out of 3522 candidates), our findings suggest that: a) there were 20 gaps related to agent-oriented methodologies; coordination, cooperation and negotiation; modelling, developing, testing and debugging; b) 24 gaps related to specific domains (recycling, dynamic evacuation, hazard management, health-care, industry, logistics and manufacturing, machine learning, ambient assisted living); and 14 gaps related to specific areas within MAS (A-Teams, dynamic MAS and mobile agents, ABMS, evolutionary MAS, and self-organizing MAS). These gaps specify lines of research where the MAS community must work to achieve the unification of the agent-oriented paradigm; as well as strengthen ties with the industry.

Towards a Hybrid MAS Organizational Model

The organizational aspects are currently getting a great attention within the multi-agent systems (MAS) community. The motivation towards this trend is finding a way to handle the increasing complexity and distribution of modern agent-based applications using higher order abstractions such as agent organizations. It is a transition from concerning the micro level (individual agents) to concerning the macro level (the whole system) to handle complexity. A large number of MAS organizational models can be found in MAS literature. Some of them adopt the ACMAS (Agent-Centered MAS) viewpoint and others adopt the OCMAS (Organizational-Centered MAS) viewpoint. Each of the ACMAS and OCMAS viewpoints has its advantages and disadvantages; therefore, combining them into a hybrid model is expected to give us the chance to take benefit of their advantages and avoid their disadvantages. This chapter presents our recent work towards the conceptual design of a hybrid MAS organizational model that combines both of the ACMAS and OCMAS viewpoints.

Adaptive cooperative formation‐containment control for networked Euler–Lagrange systems without using relative velocity information

This study is concerned with the problem of cooperative formation-containment control (CFCC) for networked Euler-Lagrange systems under a directed graph. Compared with the distributed containment control problem, CFCC problem is a new problem in the community of multi-agent systems as the control for leaders' formation is further considered. By introducing the desired relative distance between agents into the auxiliary variables, the formation control algorithms are designed firstly for multiple leaders to reach a constant configuration. Then, the distributed containment control algorithms are designed for the followers such that all followers asymptotically converge to the convex hull spanned by the leaders. In the proposed algorithms, the relative velocity information is not required for the agents' controller design. Moreover, to reduce the dependence on system global information, the control gains in the algorithms are changing adaptively using only local information. Finally, simulation results are provided to illustrate the effectiveness of the obtained theoretical results.

A socially-based distributed self-organizing algorithm for holonic multi-agent systems: Case study in a task environment

Holonic multi-agent systems (HMASs) have recently attracted many researches in multi-agent systems community. Inspired from the multi-level and self-similar structures of social and biological system, holonic multi-agent systems have been widely used to model and solve complex real-world problems. The main concern in deploying HMASs is the problem of building the hierarchical holonic structure, called holarchy, and dynamically managing it during its lifetime. The way an HMAS is organized has a great impact on its applicability and performance. This paper proposes a self-organizing algorithm to build and manage the holoic structures in multi-agent systems. This algorithm is based on the local information of the agents about other agents they can communicate with. Using common social concepts, like skills, diversity, social exchange theory, and norms in definition of the algorithm, the outcomes of this research can be used in wide ranges of distributed applications. The proposed model is extensively tested in a task allocation problem; and its performance based on various design parameters is studied. Empirical results show that the proposed model properly increases the performance of the system in terms of effectiveness and efficiency.

A CASE FOR DOMAIN-INDEPENDENT DETERMINISTIC MULTIAGENT

The notion of planning using multiple agents has been around since the very beginning of planning itself. It has been approached from various viewpoints especially in the multiagent systems community. Recently, domain-independent multiagent planning has gained more attention also in the automated planning community. In this paper, we shortly present the current state of the art, question some aspects of the research field and discuss the rising challenges.

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.

Partial observable update for subjective logic and its application for trust estimation

Subjective Logic (SL) is a type of probabilistic logic, which is suitable for reasoning about situations with uncertainty and incomplete knowledge. In recent years, SL has drawn a significant amount of attention from the multi-agent systems community as it connects beliefs and uncertainty in propositions to a rigorous statistical characterization via Dirichlet distributions. However, one serious limitation of SL is that the belief updates are done only based on completely observable evidence. This work extends SL to incorporate belief updates from partially observable evidence. Normally, the belief updates in SL presume that the current evidence for a proposition points to only one of its mutually exclusive attribute states. Instead, this work considers that the current attribute state may not be completely observable, and instead, one is only able to obtain a measurement that is statistically related to this state. In other words, the SL belief is updated based upon the likelihood that one of the attributes was observed. The paper then illustrates properties of the partial observable updates as a function of the state likelihood and illustrates the use of these likelihoods for a trust estimation application. Finally, the utility of the partial observable updates is demonstrated via various simulations including the trust estimation case.

Cooperative games and multiagent systems

AbstractForming coalitions is a generic means for cooperation: people, robots, web services, resources, firms; they can all improve their performance by joining forces. The use of coalitions has been successful in domains such as task allocations, sensor networks, and electronic marketplaces. Forming efficient coalitions requires the identification of matching synergies between different entities (finding complementary partners, or similar partners, or partners who add diversity). In addition, the different parties must negotiate a fair repartition of the worth created by the coalition. The first part of this paper is a tutorial on cooperative game theory (also called coalitional games). We then survey the different scenarios and the key issues addressed by the multiagent systems community.

A review of attacks and security approaches in open multi-agent systems

Open multi-agent systems (MASs) have growing popularity in the Multi-agent Systems community and are predicted to have many applications in future, as large scale distributed systems become more widespread. A major practical limitation to open MASs is security because the openness of such systems negates many traditional security solutions. In this paper we introduce and classify main attacks on open MASs. We then survey and analyse various security techniques in the literature and categorise them under prevention and detection approaches. Finally, we suggest which security technique is an appropriate countermeasure for which classes of attack.

Special issue: Development of service-based and agent-based computing systems

This special issue presents the best papers from theworkshops onService-OrientedComputing: Agents, Semantics and Engineering (SOCASE 2010) held in May 2010 in Toronto, Canada and the IEEE 2010 First International Workshop on Service-Oriented Computing and Multi-Agent Systems (SOCMAS 2010) held in July 2010 in Miami, Florida, USA. The goal of the workshops was to present the recent significant developments at the intersections of multi-agent systems, semantic technology, and service-oriented computing, and to promote crossfertilization of techniques. In particular, the workshops attempted to identify techniques from research on multi-agent systems and semantic technology that will have the greatest impact on automating serviceoriented application construction and management, focusing on critical challenges such as service quality assurance, reliability, and adaptability. The areas of service-oriented computing and Semantic Web services offer much of real interest to the multi-agent system community, including similarities in system architectures and provision processes, powerful tools, and the focus on many related issues including quality of service, security, and reliability. In addition, service-oriented computing and Semantic Web services offer various diverse application fields for both the concepts and methodologies of intelligent agent and multi-agent systems. Similarly, techniques developed in the multi-agent systems research community promise to have a strong impact on this fast growing technology. In particular, they enable services to be discovered and enacted across enterprise boundaries. If an organisation bases its success on services provided by others, then it must be able to trust that the services will perform as promised, whenever needed. Researchers in multi-agent systems have investigated such trust mechanisms.

Simulation in Agent-Oriented Software Engineering: The SODA case study

The key role of simulation in the engineering of complex multiagent systems (MAS) is today generally acknowledged in the MAS community. However, the adoption of simulation in state-of-the-art Agent-Oriented Software Engineering (AOSE) methodologies is still incomplete at its best.In this paper we present a simulation-based approach to MAS engineering and discuss its integration within AOSE methodologies. Integration is first discussed in general by adopting standard method engineering techniques, then detailed by means of a case study—that is, integrating simulation in SODA.

Dependence in games and dependence games

In the multi-agent systems community, dependence theory and game theory are often presented as two alternative perspectives on the analysis of agent interaction. The paper presents a formal analysis of a notion of dependence between players, given in terms of standard game-theoretic notions of rationality such as dominant strategy and best response. This brings the notion of dependence within the realm of game theory providing it with the sort of mathematical foundations which still lacks. Concretely, the paper presents two results: first, it shows how the proposed notion of dependence allows for an elegant characterization of a property of reciprocity for outcomes in strategic games; and second, it shows how the notion can be used to define new classes of coalitional games, where coalitions can force outcomes only in the presence of reciprocal dependencies.

Environment in agent-oriented software engineering methodologies

The key role of environment as a first-class abstraction in the engineering of MAS (multi-agent systems) is today generally acknowledged in the MAS community. However, the support for the notion of environment in today AOSE (agent-oriented software e

Industry traction for MAS technology: would a rose by any other name smell as sweet?

The Multi-Agent Systems (MAS) community faces a crisis that many are unwilling to acknowledge. We are all acutely aware of how MAS concepts and technologies seem to have failed to gain significant industry traction (a few notable exceptions aside). Other research communities (e.g., the service-oriented computing community) have gained far greater industry acceptance for their outputs, in far shorter time. But the crisis stems from an even greater threat: the co-opting of agents/MAS concepts in other research communities, particularly the service-oriented computing community. This position paper argues that the MAS community must respond to this challenge via a closer engagement with the service-oriented computing community, by using the notion of services as an implementation vehicle for MAS technology, by better articulating the industry value proposition of MAS technology in software engineering terms and by paying attention to nomenclature.

How to get multi-agent systems accepted in industry?

With many researchers in the Multi-Agent System (MAS) community, we share the opinion that too much of the quality and relevant research in the area of MAS is underrepresented in the development of complex distributed systems in practice today. In our experience, a Babylonic mismatch is a crucial factor in this fact – research on MAS profiles itself as an isolated community and, as such, may create artificial thresholds to convince mainstream software developers of its merits. We argue that integrating the concepts and techniques from agent-based software engineering within mainstream software engineering provides opportunities to amplify the industrial adoption of MAS. To ground this position, we discuss MAS engineering from the perspective of the software engineering area that we are most familiar with: software architecture.