Governing Black-Box Agents in Competitive Multi-Agent Systems

Competitive Multi-Agent Systems (MAS) are inherently hard to control due to agent autonomy and strategic behavior, which is particularly problematic when there are system-level objectives to be achieved or specific environmental states to be avoided. Existing methods mostly assume specific knowledge about agent preferences, utilities and strategies, neglecting the fact that actions are not always directly linked to genuine agent preferences, but can also reflect anticipated competitor behavior, be a concession to a superior adversary or simply be intended to mislead other agents. This assumption both reduces applicability to real-world systems and opens room for manipulation. We therefore propose a new governance approach for Multi-Attribute MAS which relies exclusively on publicly observable actions and transitions, and uses the acquired knowledge to purposefully restrict action spaces, thereby achieving the system's objectives while preserving a high level of autonomy for the agents.

Multi-agent systems can solve scientific issues related to complex systems that are difficult or impossible for a single agent to solve through mutual collaboration and cooperation optimization. In a multi-agent system, agents with a certain degree of autonomy generate complex interactions due to the correlation and coordination, which is manifested as cooperative/competitive behavior. This survey focuses on multi-agent cooperative optimization and cooperative/non-cooperative games. Starting from cooperative optimization, the studies on distributed optimization and federated optimization are summarized. The survey mainly focuses on distributed online optimization and its application in privacy protection, and overviews federated optimization from the perspective of privacy protection mechanisms. Then, cooperative games and non-cooperative games are introduced to expand the cooperative optimization problems from two aspects of minimizing global costs and minimizing individual costs, respectively. Multi-agent cooperative and non-cooperative behaviors are modeled by games from both static and dynamic aspects, according to whether each player can make decisions based on the information of other players. Finally, future directions for cooperative optimization, cooperative/non-cooperative games, and their applications are discussed.

Restriction-based governance has recently been proposed as an alternative to reward shaping for achieving system-level goals in competitive multi-agent systems. In this work, we apply these two approaches to the domain of traffic management, specifically investigating their efficacy and fairness. Our results show that edge restrictions in congested traffic networks are superior to dynamic pricing with regard to equity (i.e., equal treatment of agents) while achieving comparable travel-time improvements. We argue that the former metric, as an adequate proxy for fairness, can be crucial for the quality and acceptance of a governance scheme, particularly when human agents are affected.

The effects of multiple-pedagogical agents on learners’ academic success, motivation, and cognitive load

The research focuses on methods and software tools for decision support for the organization of distributed production of unmanned aerial vehicles (UAVs) based on digital twin and multi-agent systems technologies. The purpose of the work is to provide a critical review of existing approaches to the organization of distributed production of UAVs, as complex technical objects, and to substantiate an approach to building a multi-agent decision support system for the synthesis of a rational production structure based on digital twins of participants in production cooperation. The tasks include: classifying unmanned aerial vehicles according to the main features that determine their production specifics; analysing modern scientific publications in the field of distributed production informatization, multi-agent systems, digital twins and intelligent decision support; identifying the limitations of existing solutions; and proposing an approach to the organization of distributed production of UAVs using digital twins in the form of intelligent agents. The applied methods are based on system and classification analysis, generalization of modern scientific approaches to the organization of production, decision-making theory, the concept of multi-agent systems, digital twin technologies, production logistics and artificial intelligence. The following results were obtained: UAVs were classified by mass, purpose and level of autonomy, which made it possible to determine the specifics of their production and life cycle management. An analysis of scientific publications was performed, which showed the active development of areas related to the use of multi-agent systems, digital twins and artificial intelligence to address complex tasks in UAV life cycle management. It was established, in particular, that the works of recent years consider the issues of cooperative UAV production, the use of deep learning algorithms for decision-making in dynamic environments and the integration of digital twins into relevant production systems. At the same time, a number of limitations on the implementation of existing solutions were identified: insufficient formalization of business processes between production participants, weak integration of digital twins into production logistics, and the lack of self-organization mechanisms. An approach to solving the problem of effective organization of distributed UAV production is proposed. The architecture of a multi-agent decision support system for distributed UAV production is developed, which includes modules for coordination, self-organization, and the evaluation of the effectiveness of production configuration options, considering the initial risk tolerance indicator. Conclusions. The scientific novelty of the results obtained is as follows: an approach to solving the problem of effective organization of distributed UAV production is proposed, which, unlike existing ones, is based on the representation of production participants as digital twins in the form of intelligent agents, this enables the generation of multiple options and the subsequent selection of the most rational configuration for organizing distributed UAV production in terms of business process efficiency and the reduction of production risks. The research results create a basis for further integration of multi-agent technologies into real-world manufacturing platforms and the expansion of their application to other types of complex machinery.

An adaptive evaluation mechanism for online traders

We introduce and study SL[F], a quantitative extension of SL (Strategy Logic), one of the most natural and expressive logics describing strategic behaviours. The satisfaction value of an SL[F] formula is a real value in [0,1], reflecting ``how much'' or ``how well'' the strategic on-going objectives of the underlying agents are satisfied. We demonstrate the applications of SL[F] in quantitative reasoning about multi-agent systems, by showing how it can express concepts of stability in multi-agent systems, and how it generalises some fuzzy temporal logics. We also provide a model-checking algorithm for ourlogic, based on a quantitative extension of Quantified CTL*.

Intrusion detection systems play an important role in preventing attacks which have been increased rapidly due to the dependence on network and Internet connectivity. Deep learning algorithms are promising techniques, which have been used in many classification problems. In the same way, multi-agent systems become a new useful approach in intrusion detection field. In this paper, we propose a deep learning-based multi-agent system for intrusion detection which combines the desired features of multi-agent system approach with the precision of deep learning algorithms. Therefore, we created a number of autonomous, intelligent and adaptive agents that implanted three algorithms, namely autoencoder, multilayer perceptron and k-nearest neighbors. Autoencoder is used as features reduction tool, and multilayer perceptron and k-nearest neighbors are used as classifiers. The performance of our model is compared against traditional machine learning approaches and other multi-agent system-based systems. The experiments have shown that our hybrid distributed intrusion detection system achieves the detection with better accuracy rate and it reduces considerably the time of detection.

A paradigm known as multi-agent systems (MAS) combines several independent agents to accomplish intricate tasks. The creation of such systems presents difficulties with control flow, specialization, and modularity. With an emphasis on Lang-Graphbased implementation, this study examines various MAS topologies, such as network-based, supervisor-led, and hierarchical models. According to the results, MAS enhances performance and modularity while requiring careful inter-agent communication strategy design. This paper provides a thorough analysis of the architecture of Multi-Agent Systems (MAS), emphasizing its fundamental ideas, constituent parts, and uses. In the advanced distributed computing paradigm known as MAS, autonomous agents work together to resolve challenging issues. In “order to shed light on the possibilities of MAS in a number of fields, including as robotics, e-commerce, transportation, and healthcare, this study examines agent definitions, system architecture, communication protocols, decision-making methods, and practical applications. Autonomous agents are integrated in Multi-Agent Systems (MAS) to address complex issues, although control flow, specialization, and modularity present difficulties. This paper examines three topologies-network-based, supervisorled, and hierarchical-detailing inter-agent communication protocols and coordination mechanisms with an emphasis on Lang-Graph-based implementations. Each architecture is demonstrated with sample code, and the findings demonstrate that, with well-thought-out communication mechanisms, MAS improves performance and modularity. Applications include e-commerce, healthcare, transportation, and robots.

Environmental concerns and the shortage in the fossil fuel reserves have been potentiating the growth and globalization of distributed generation. Another resource that has been increasing its importance is the demand response, which is used to change consumers' consumption profile, helping to reduce peak demand. Aiming to support small players' participation in demand response events, the Curtailment Service Provider emerged. This player works as an aggregator for demand response events. The control of small and medium players which act in smart grid and micro grid environments is enhanced with a multi-agent system with artificial intelligence techniques - the MASGriP (Multi-Agent Smart Grid Platform). Using strategic behaviours in each player, this system simulates the profile of real players by using software agents. This paper shows the importance of modeling these behaviours for studying this type of scenarios. A case study with three examples shows the differences between each player and the best behaviour in order to achieve the higher profit in each situation.

The issue of creating an information security system is very relevant in the world today. One of the urgent tasks is to solve the issues of effective protection of information from both external and internal threats through the creation and implementation of information security management systems in automated systems of enterprises, which, among other things, requires the formalization of the task of protecting information for its subsequent implementation by software and other means. Now there are security analysis systems, for example, that examine the security elements settings of workstations and servers operating systems, analyze the network topology, look for unprotected network connections, examine the settings of firewalls. The disadvantage of these systems is that they are not suitable for monitoring large volumes of network traffic. The solution to this problem is the use of monitoring tools capable of analyzing large amounts of data in real time. Therefore, a significant place in the article is given to the review of developments based on artificial intelligence technologies, namely multi-agent systems, review of information security models, threat risk assessment in automated systems. The functional architecture of the information security management system based on a multi-agent system has been proposed to search in real time for information security optimal solutions through the selection of such coalitions of protection mechanisms agents that will allow to build the optimal protection of the automated system according to the selected criteria. The model with complete overlapping of threats has been substantiated and adopted as a basis, which allows to analyze the overall situation and choose strategically important decisions directly during the organization of information security. The essence of of multi-agent systems functioning that implement a decentralized control system based on the work of autonomous agents that can be implemented programmatically has been revealed. The role of threat agents, resource agents, agents of protection mechanisms and their functional purpose have been defined. The problem of searching a set of protection mechanisms agents coalition for the current state of the automated system as a problem of optimal search by the criterion of protection cost, taking into account the value of information, has been generalized. Due to the modularity of the multi-agent system, the further work will be aimed at detailing its components and perfection.

The Distributed Constraint Optimization Problem (DCOP) has been studied as a fundamental optimization problem that represents various problems on multiagent systems. We focus on the asymmetric DCOPs where each objective function is differently defined as an evaluation of an agent. This class of problems is studied as a multi-objective problem for the preferences of individual agents. In this work, we investigate the possibility of a solution framework based on relaxation methods as a scalable and inexact solution approach for this class of problems. We address a bottleneck problem that minimizes the worst-case cost value. As the first study, we apply a penalty method to the minimization problems of the maximum cost values.

The purpose of this chapter is to describe the application of the problem solving ontology component of HCVM in the area of face detection and annotation. computer-based aspects of an embedded feature-based multi-agent face detection and annotation system are described in this chapter. Unlike similar feature-based systems that start from searching for facial organs in the images and group them to find faces, the system solves the problem using the problem solving agents of the HCVM in a top-down fashion. The embedded mtilti-agent face detection system makes hypotheses of face locations and seeks evidences to verify them. The five problem solving agents follow a coarse grain to fine grain methodology and work on various features of color images. The coarse to fine top-down methodology is more akin to human perceptions than a bottom-up approach. The higher level problem solving agents like decomposition work as background knowledge for the control and decision agents resulting in improvement of the accuracy and speed of detection. It is also distinct from many other top-down systems in that no re-scaling of the images is needed in the searching process. To improve the detection rate, a unique iterative region-partitioning algorithm is developed in this multi-agent system. The problem solving agents model aspects related to skin-tone region segmentation, noise reduction, candidate face regions location and facial feature extraction/face detection, and annotation. The detection algorithm is invariant to scale and rotation to some degree. The performance of the multi-agent system has been tested with a large number of images and the results obtained thus far are very encouraging.

Welcome to the Eighth Annual Workshop on Evolutionary Computation and Multi-Agent Systems and Simulation (ECoMASS 2014)! Evolutionary computation (EC) and multi-agent systems and simulation (MASS) both involve populations of agents. EC is a learning technique by which populations of individual agents adapt according to the selection pressures exerted by an environment; MASS seeks to understand how to coordinate the actions of a population of (possibly selfish) autonomous agents that share an environment so that some outcome is achieved. Both EC and MASS have top-down and bottom up features. For example, some aspects of multi-agent system engineering (e.g., mechanism design) are concerned with how top-down structure can constrain or influence individual decisions. Similarly, most work in EC is concerned with how to engineer selective pressures to drive the evolution of individual behavior towards some desired goal. Multi-agent simulation (also called agent-based modeling) addresses the bottom-up issue of how collective behavior emerges from individual action. Likewise, the study of evolutionary dynamics within EC (for example in coevolution) often considers how population-level phenomena emerge from individual-level interactions. Thus, at a high level, we may view EC and MASS as examining and utilizing analogous processes. It is therefore natural to consider how knowledge gained within EC may be relevant to MASS, and vice versa; indeed, applications and techniques from one field have often made use of technologies and algorithms from the other field. Studying EC and MASS in combination is warranted and has the potential to contribute to both fields. The ECoMASS Workshop at GECCO has a successful history as a forum for exploring precisely this intersection, and we are looking forward to another year of stimulating discussion, bringing together experts as well as novices in both areas, to engage in dialogue about their work. This year's participants bring a variety of topics for discussion, including migratory flows, the beating of the heart, embedded systems, pelotons and cyclists, and flocking behavior. This year, we will also be adding a research slam. We will be inviting presenters who are presenting in the general conference to give us a quick presentation about their work. We hope that this will further encourage additional lively discussion about a wide range of topics. We also encourage presenters to demonstrate their software live during a breakout session, enabling additional opportunities for one-on-one dialogue and discussion of actual running systems. Through these additions to the workshop, we hope to continue to innovate, and develop the conversations around the interesting intersection of evolutionary computation and multi-agent systems and simulations!

The paper explores the development of a multi-agent scalable system (SMA) concept for monitoring and predicting environmental processes, addressing the challenges related to the organization of distributed computing processes and the application of knowledge-based models, machine learning and artificial intelligence. SMA offers an innovative solution based on autonomous agents that perceive the environment, collaborate with each other, with the data storage center and implement advanced computing models. The technical and technological development of the system provides for the use of a set of smart sensors and devices, such as the ESP32, which integrate acquisition, data preprocessing and network communication services. The agent functional diagram and the algorithm for collaboration between agents and the data storage center were developed. The process of predicting environmental events is based on the application of neural network models. It was analyzed the carbon dioxide pollution scenario and its evolution to validate the proposed concept. The paper demonstrates that multi-agent systems can provide efficient and scalable solutions for managing environmental problems, to help to prevent crises and to support environmental protection policies.