Multi-agent System Model Research Articles

Hierarchical Behavior Model for Multi-Agent System with Evasion Capabilities and Dynamic Memory

The behavior of an agent may be simple or complex depending on its role. Behavioral simulation using agents can have multiple approaches that have different advantages and disadvantages. By combining different behaviors in a hierarchical model, situational inefficiencies can be compensated. This paper proposes a behavioral hierarchy model that combines different mechanisms in behavior plans. The study simulates the social behavior in an office environment during an emergency using collision avoidance, negotiation, conflict solution, and path-planning mechanisms in the same multi-agent model to find their effects and the efficiency of the combinational setups. Independent agents were designed to have memory expansion, pathfinding, and searching capabilities, and the ability to exchange information among themselves and perform evasive actions to find a way out of congestion and conflict. The designed model allows us to modify the behavioral hierarchy and action order of agents during evacuation scenarios. Moreover, each agent behavior can be enabled or disabled separately. The effects of these capabilities on escape performance were measured in terms of time required for evacuation and evacuation ratio. Test results prove that all mechanisms in the proposed model have characteristics that fit each other well in situations where different hierarchies are needed. Dynamic memory management (DMM), together with a hierarchical behavior plan, achieved a performance improvement of 23.14% in escape time without providing agents with any initial environmental information.

Agent-Based Distributed Planning and Coordination for Resilient Airport Surface Movement Operations

Airport surface movement operations are complex processes with many types of adverse events which require resilient, safe, and efficient responses. One regularly occurring adverse event is that of runway reconfiguration. Agent-based distributed planning and coordination has shown promising results in controlling operations in complex systems, especially during disturbances. In contrast to the centralised approaches, distributed planning is performed by several agents, which coordinate plans with each other. This research evaluates the contribution of agent-based distributed planning and coordination to the resilience of airport surface movement operations when runway reconfigurations occur. An autonomous Multi-Agent System (MAS) model was created based on the layout and airport surface movement operations of Schiphol Airport in the Netherlands. Within the MAS model, three distributed planning and coordination mechanisms were incorporated, based on the Conflict-Based Search (CBS) Multi-Agent Path Finding (MAPF) algorithm and adaptive highways. MAS simulations were run based on eight days of real-world operational data from Schiphol Airport and the results of the autonomous MAS simulations were compared to the performance of the real-world human operated system. The MAS results show that the distributed planning and coordination mechanisms were effective in contributing to the resilient behaviour of the airport surface movement operations, closely following the real-world behaviour, and sometimes even surpassing it. In particular, the mechanisms were found to contribute to more resilient behaviour than the real-world when considering the taxi time after runway reconfiguration events. Finally, the highway included distributed planning and coordination mechanisms contributed to the most resilient behaviour of the airport surface movement operations.

A multi-agent system for distributed smartphone sensing cycling in smart cities

PurposeThe purpose of this paper is to propose a distributed smartphone sensing-enabled system, which assumes an intelligent transport signaling (ITS) infrastructure that operates traffic lights in a smart city (SC). The system is able to handle priorities between groups of cyclists (crowd-cycling) and traffic when approaching traffic lights at road junctions.Design/methodology/approachThe system takes into consideration normal probability density function (PDF) and analytics computed for a certain group of cyclists (i.e. crowd-cycling). An inference model is built based on real-time spatiotemporal data of the cyclists. As the system is highly distributed – both physically (i.e. location of the cyclists) and logically (i.e. different threads), the problem is treated under the umbrella of multi-agent systems (MAS) modeling. The proposed model is experimentally evaluated by incorporating a real GPS trace data set from the SC of Melbourne, Australia. The MAS model is applied to the data set according to the quantitative and qualitative criteria adopted. Cyclists’ satisfaction (CS) is defined as a function, which measures the satisfaction of the cyclists. This is the case where the cyclists wait the least amount of time at traffic lights and move as fast as they can toward their destination. ITS system satisfaction (SS) is defined as a function that measures the satisfaction of the ITS system. This is the case where the system serves the maximum number of cyclists with the fewest transitions between the lights. Smart city satisfaction (SCS) is defined as a function that measures the overall satisfaction of the cyclists and the ITS system in the SC based on CS and SS. SCS defines three SC policies (SCP), namely, CS is maximum and SS is minimum then the SC is cyclist-friendly (SCP1), CS is average and SS is average then the SC is equally cyclist and ITS system friendly (SCP2) and CS is minimum and SS is maximum then the SC is ITS system friendly (SCP3).FindingsResults are promising toward the integration of the proposed system with contemporary SCs, as the stakeholders are able to choose between the proposed SCPs according to the SC infrastructure. More specifically, cyclist-friendly SCs can adopt SCP1, SCs that treat cyclists and ITS equally can adopt SCP2 and ITS friendly SCs can adopt SCP3.Originality/valueThe proposed approach uses internet connectivity available in modern smartphones, which provide users control over the data they provide to us, to obviate the installation of additional sensing infrastructure. It extends related study by assuming an ITS system, which turns traffic lights green by considering the normal PDF and the analytics computed for a certain group of cyclists. The inference model is built based on the real-time spatiotemporal data of the cyclists. As the system is highly distributed – both physically (i.e. location of the cyclists) and logically (i.e. different threads), the system is treated under the umbrella of MAS. MAS has been used in the literature to model complex systems by incorporating intelligent agents. In this study, the authors treat agents as proxy threads running in the cloud, as they require high computation power not available to smartphones.

Generative Attention Networks for Multi-Agent Behavioral Modeling

Understanding and modeling behavior of multi-agent systems is a central step for artificial intelligence. Here we present a deep generative model which captures behavior generating process of multi-agent systems, supports accurate predictions and inference, infers how agents interact in a complex system, as well as identifies agent groups and interaction types. Built upon advances in deep generative models and a novel attention mechanism, our model can learn interactions in highly heterogeneous systems with linear complexity in the number of agents. We apply this model to three multi-agent systems in different domains and evaluate performance on a diverse set of tasks including behavior prediction, interaction analysis and system identification. Experimental results demonstrate its ability to model multi-agent systems, yielding improved performance over competitive baselines. We also show the model can successfully identify agent groups and interaction types in these systems. Our model offers new opportunities to predict complex multi-agent behaviors and takes a step forward in understanding interactions in multi-agent systems.

Bisimulations for verifying strategic abilities with an application to the ThreeBallot voting protocol

We propose a notion of alternating bisimulation for strategic abilities under imperfect information. The bisimulation preserves formulas of ATL⁎ for both the objective and subjective variants of the state-based semantics with imperfect information, which are commonly used in the modeling and verification of multi-agent systems. Furthermore, we apply the theoretical result to the verification of coercion-resistance in the ThreeBallot voting system, a voting protocol that does not use cryptography. In particular, we show that natural simplifications of an initial model of the protocol are in fact bisimulations of the original model, and therefore satisfy the same ATL⁎ properties, including coercion-resistance. These simplifications allow the model-checking tool MCMAS to terminate on models with a larger number of voters and candidates, compared with the initial model.

Multi-player games with LDL goals over finite traces

Linear Dynamic Logic on finite traces (LDLF) is a powerful logic for reasoning about the behaviour of concurrent and multi-agent systems. In this paper, we investigate techniques for both the characterisation and verification of equilibria in multi-player games with goals/objectives expressed using logics based on LDLF. This study builds upon a generalisation of Boolean games, a logic-based game model of multi-agent systems where players have goals succinctly represented in a logical way. Because LDLF goals are considered, in the settings we study—Reactive Modules games and iterated Boolean games with goals over finite traces—players' goals can be defined to be regular properties while achieved in a finite, but arbitrarily large, trace. In particular, using alternating automata, the paper investigates automata-theoretic approaches to the characterisation and verification of (pure strategy Nash) equilibria, shows that the set of Nash equilibria in multi-player games with LDLF objectives is regular, and provides complexity results for the associated automata constructions.

An intelligent real-time scheduler for out-patient clinics: A multi-agent system model.

Scheduling of resources and patients are crucial in outpatient clinics, particularly when the patient demand is high and patient arrivals are random. Generally, outpatient clinic systems are push systems where scheduling is based on average demand prediction and is considered for long term (monthly or bimonthly). Often, planning and actual scenario vary due to uncertainty and variability in demand and this mismatch results in prolonged waiting times and under-utilization of resources. In this article, we model an outpatient clinics as a multi-agent system and propose an intelligent real-time scheduler that schedules patients and resources based on the actual status of departments. Two algorithms are implemented: one for resource scheduling that is based on predictive demand and the other is patient scheduling which performs path optimization depending on the actual status of departments. In order to match resources with stochastic demand, a coordination mechanism is developed that reschedules the resources in the outpatient clinics in real time through auction-bidding procedures. First, a simulation study of intelligent real-time scheduler is carried out followed by implementation of the same in an outpatient clinic of Aravind Eye Hospital, Madurai, India. This hospital has huge patient demand and the patient arrivals are random. The results show that the intelligent real-time scheduler improved the performance measures like waiting time, cycle time, and utilization significantly compared to scheduling of resources and patients in isolation. By scheduling resources and patients, based on system status and demand, the outpatient clinic system becomes a pull system. This scheduler transforms outpatient clinics from open loop system to closed-loop system.

Finite horizon discrete models for multi-agent control systems with coupled dynamics

The goal of this paper is to obtain online abstractions for coupled multi-agent systems in a decentralized manner. A discrete model which captures the motion capabilities of each agent is derived over a bounded time-horizon, by discretizing a corresponding overapproximation of the agent’s reachable states. The individual abstractions’ composition provides a correct representation of the coupled continuous system over the horizon and renders the approach appropriate for control synthesis under high-level specifications which are assigned to the agents over this time window. Sufficient conditions are also provided for the space and time discretization to guarantee the derivation of deterministic abstractions with tunable transition capabilities.

A Hybrid Event-Triggered Approach to Consensus of Multiagent Systems With Disturbances

This article investigates the event-triggered consensus problem for multiagent systems (MAS) with external disturbances. A hybrid event-triggering mechanism (ETM) is proposed, which utilizes time regularization to force the minimum interevent time to be strictly positive, such that Zeno freeness is ensured in the presence of disturbances. Model-based control protocol is used to avoid continuous communication between neighboring agents. With the proposed hybrid ETM, the closed-loop MAS contains both flow dynamics and jumping dynamics, which cannot be described by the existing MAS models. By means of hybrid system framework and auxiliary variables, a hybrid model of the MAS is constructed to describe the flow dynamics and jumping dynamics. Based on this model, the Lypuanov-based consensus analysis and hybrid ETM design are developed, and the finite-gain L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> stability property is guaranteed. An example of spacecraft formation is provided to show the effectiveness of the proposed approach.

Study on Master Slave Interaction Model Based on Stackelberg Game in Distributed Environment

In view of the problems such as low efficiency, difficulty in resolving local conflicts and lack of practical application scenarios, existing in the interaction model of multi-agent systems in a distributed environment, a multi-master multi-slave interaction model was designed based on the Stackelberg game, which is applied to the interaction game problem between the controller and the participant in the command and control process. Through optimizing the Stackelberg game model and multi-attribute decision-making, the multi-master, multi-slave, multi-agent system of the Stackelberg game was designed, and the closed loop problem under the Stackelberg game is solved for dimension reduction and optimal function value. Finally, through the numerical derivation simulation and the training results of related system data, the high efficiency and strong robustness of the model were verified from multiple perspectives, and this model algorithm was proved to be true and highly efficient.

ПРИМЕНЕНИЕ ВЕРОЯТНОСТНЫХ И ВРЕМЕННЫХ АВТОМАТОВ В ПРОГРАММАХ УПРАВЛЕНИЯ МНОГОАГЕНТНЫХ СИСТЕМ

Classical models of the time and probabilistic automaton constructed using the theory of Markov chains - the system "automaton-random environment" described by the Markov chain are considered. The uncertainty of the probabilistic automaton can be determined by various reasons, including the nondeterminism of the rules of state change of the simulated system, which is one of the most important obstacles in the modeling of multi-agent systems. The developed model is intended for the control program of multi-agent systems, takes into account the stochastic behavior of the environment when used in multi-agent systems. The key differences between the classical stationary random environment and multi-agent systems are presented, including the rationality of agent behavior and the need for cooperation or counteraction. The model does not provide trainable algorithms, their application requires a large sample for training, not all multi-agent systems have enough time to set the necessary statistics and behavior adaptation. Approaches to building the architecture of the agent control program in a multi-agent environment based on the theories of time and probabilistic automata are described, a format for storing models in file storage is proposed. The proposed architecture takes into account the possibility of changing the state of time automata based on the state of probabilistic automata and vice versa. The generalized algorithm of the control program functioning is described. The proposed algorithm makes it possible to implement control programs for multi-agent systems in the case of a finite number of agent states and uses flexible control mechanisms taken from the time machine model. The main results and proposals for further research, including the construction of hierarchical models are presented.

Intelligent agent for formal modelling of temporal multi-agent systems

Abstract Software systems are becoming complex and dynamic with the passage of time, and to provide better fault tolerance and resource management they need to have the ability of self-adaptation. Multi-agent systems paradigm is an active area of research for modeling real-time systems. In this research, we have proposed a new agent named SA-ARTIS-agent, which is designed to work in hard real-time temporal constraints with the ability of self-adaptation. This agent can be used for the formal modeling of any self-adaptive real-time multi-agent system. Our agent integrates the MAPE-K feedback loop with ARTIS agent for the provision of self-adaptation. For an unambiguous description, we formally specify our SA-ARTIS-agent using Time-Communicating Object-Z (TCOZ) language. The objective of this research is to provide an intelligent agent with self-adaptive abilities for the execution of tasks with temporal constraints. Previous works in this domain have used Z language which is not expressive to model the distributed communication process of agents. The novelty of our work is that we specified the non-terminating behavior of agents using active class concept of TCOZ and expressed the distributed communication among agents. For communication between active entities, channel communication mechanism of TCOZ is utilized. We demonstrate the effectiveness of the proposed agent using a real-time case study of traffic monitoring system.

Доверительная модель информационной безопасности среды многоагентных робототехнических систем

The capabilities of robotic systems are growing rapidly. The ways of operating such systems are increasing every day. Sensor networks, the Internet of Things, cyber-physical systems have similar properties as multi-agent robotic systems, which allows us to consider the approaches and methods of ensuring security used in them. Purpose of the article: improving the security of multi-agent systems in an uncontrolled environment, developing methods for assessing trust of the environment. Research methods: analysis of existing threat models for multi-agent robotic systems, as well as systems with similar properties: cyber-physical systems and the Internet of things. Analysis of the research results of existing approaches to ensuring the security of multi-agent systems. The results: an analysis of security threats and existing methods of ensuring security for robotic multi-agent systems, as well as systems with similar properties, was carried out. An extension for security methods based on trust and reputation has been developed, taking into account the operating environment of the system as part of information interaction. A method for localizing the subject of an external intruder in the environment is proposed.

A Methodological Approach to the Assessment of the Impact of Digital Technologies Development in Energy Industry on Electricity Price and Demand in a Region

The relevance of this study is due to the importance of assessing the prospective dynamics and structure of demand for energy carriers when developing and making strategic decisions in the field of energy and economic security of the country and its regions. The advance of digital technology redefines the properties of electric power supply systems, erases the boundary between electric power producers and consumers, and impacts the formation of electricity price and demand in the region. This study presents a method of electricity costing in the regional power system, which serves as an integral part of the approach to assessing the impact of intelligent systems development on the demand for electricity in the region. The approach is unique in that it simulates the behavior of electricity consumers and producers of various types as they pursue their own interests and assesses the impact of this behavior on the demand and price of electricity in the regional power system. Determining the cost of electricity in the system is based on the consistent alignment of the required amount of electricity consumption with the capabilities of producers seeking to achieve their best economic performance. Each producer is described as an optimization model, which is a standalone agent in a multi-agent power system model.

Adaptive Negotiation Model for Human-Machine Interaction on Decision Level

Abstract Our work is a contribution to automation design for human-machine cooperation with explicit emancipated cooperative decision making. We propose an adaptive negotiation framework as a model for human-machine interaction on decision level. This expands modeling of human-machine cooperation, starting at the stabilization and trajectory level with approaches such as shared control, towards higher levels of interaction as guidance and navigation. In essence, the framework extends the well-known basic negotiation model of multi-agent systems by an explicit adaptation of the agent’s negotiation behavior. The adaptation is based on an opponent model using a Bayesian learning approach. An exemplary implementation for the application of human-automation interaction in autonomous driving is introduced. First results prove the high flexibility of the framework to model human negotiation behavior.

Emergent dynamics of an orientation flocking model for multi-agent system

We study the orientation flocking for the deterministic counterpart of a stochastic agent-based model introduced by Degond, Frouvelle and Merino-Aceituno in 2017, where the orientation is defined as a \begin{document}$ {\rm SO}(3) $\end{document} matrix. Their proposed model can be reduced to the other collective dynamics models such as the Lohe matrix model and the Viscek-type model as special cases. In this work, we study the emergent dynamics of the orientation flocking model in two frameworks. First, we present sufficient conditions leading to the orientation flocking when the natural frequency matrices are identical. To be precise, we prove that all orientation matrices asymptotically converge to the common one, and the spatial position diameter remains uniformly bounded. Second, we show the emergence of orientation-locked states for non-identical natural frequency matrices, that is, the difference of any two orientation matrices tends to the definite constant matrix. On the other hand, we establish the finite-in-time stability with respect to initial data of the proposed orientation flocking model. We also present the numerical results consistent with our rigorous analysis. Our work remains valid even for dimensions greater than three.

Second-Order Consensus of Hybrid Multi-Agent Systems With Unknown Disturbances Via Sliding Mode Control

In this paper, we consider the second-order consensus problem of hybrid multi-agent systems with unknown disturbances by using sliding mode control under the leader-follower network. First, the hybrid multi-agent system model with disturbances and nonlinear term is proposed, which is composed of continuous-time dynamic agents and discrete-time dynamic agents. Second, the definition of the second-order consensus of hybrid multi-agent system is given. Then, we assume that the interaction among all agents happens in sampling time and each continuous-time dynamic agent can observe its own states in real time. Based on the equivalent approaching law and the states information among agents, the sliding mode control protocols are designed to achieve the second-order consensus of the hybrid multi-agent system. Some sufficient conditions are given for solving the second-order consensus under the sliding mode control protocols. Finally, some simulations are also given to illustrate the validity of the proposed method.

Модель агента для многоагентной организационной системы: экономический аспект

Modeling the behavior of social and economic systems plays an important role in various fields of knowledge. An important task is the prediction of the behavior of such systems and their states in the future, for the adoption of proactive and corrective management decisions. Despite the constant development of game theory and other disciplines of a socio-economic (cybernetic) profile, the issue of a generalized formalized description of such systems consisting of active elements remains relevant. Significant difficulties arise with the formal description of any systems that contain active elements (people), due to the human factor and the lack of reliable knowledge about the features of the human brain. Many theories, in particular the theory of management of organizational systems, use the idea of an “economic” person as a hypothesis, whose preferences can be described by the objective cost and effect functions, which is typical for “western” scientific trends. In most cases, the developers of such theories do not propose anything concrete, but merely shift the responsibility to decision makers who are invited to build all the models on their own. It has been stated that so far many social and economic processes are considered as strategic games and zero-sum games, which leaves its mark on the decision-making process. An agent model for multi-agent organizational systems has been formed, which can be used to develop planning, incentive mechanisms that are focused on the needs and wishes of work performers. Elements of the agent model under consideration are implemented in the system for monitoring the effectiveness of the activity of the Siberian State Industrial University.

NorMAS-ML: Supporting the Modeling of Normative Multi-agent Systems

Context: A normative multi-agent system (NMAS) is composed of agents that their behavior is regulated by norms. The modeling of those elements (agents and norms) together at design time can be a good way for a complete understanding of their structure and behavior. Multi-agent system modeling language (MAS-ML) supports the representation of NMAS entities, but the support for concepts related to norms is somewhat limited. MAS-ML is founded in taming agents and objects (TAO) framework and has a support tool called the MAS-ML tool. Goal: The present work aims to present a UML-based modeling language called normative multi-agent system (NorMAS-ML) able to model the MAS main entities along with the static normative elements. Method: We extend the TAO adding normative concepts and spread out those concepts in two syntaxes of MAS-ML. Either abstract, adding or updating metaclasses and stereotypes or concrete, defining new graphic elements for representing the elements defined in the abstract syntax. Besides, we evolve the MAS-ML tool, considering the extension of MAS-ML by the model-driven approach. Results: NorMAS-ML, the new version of MAS-ML, allows a complete view of NMAS entities and has a support tool called NorMAS-ML tool. Beyond the definition of NorMAS-ML and its tool, we generate a new static diagram called “norm diagram” supported by the NorMAS-ML tool. In order to illustrate the syntax of NorMAS-ML, the entities of a conference management system and its norms are modeled jointly. Conclusion: NorMAS-ML can help software designers (i) to understand the properties and behavior of NMAS entities and (ii) to provide a software modeling following the stakeholders’ need and less complex for the development phase.

Multiagent approach for consensus control in the energy internet network

Energy internet network is an energy supply and demand system which based on the power grid, that can make a variety of energy interoperability, allowing all constituent elements to exchange information and interaction. However, any anomaly/failure in a part of the network can quickly spread over the network and lead to new unpredicted failures. Hence, timely cyber-attacks detection and mitigation is crucial. In this study, for timely detection, firstly propose a distributed state estimator assuming regular system operation, that achieves near-optimal performance based on the local Kalman filters and with the exchange of necessary information between local nodes over EI network. To enhance the security, consensus control approach can provide fault tolerance and timely attack mitigation and further recovery, thereby providing the EI with strong security in distributed. In this paper, therefore, a distributed Kalman filter with consensus control approach for multiagent systems (MAS) was proposed to investigate the energy internet recovered with links failure by cyber-attacks. The proposed algorithm is verified in the energy internet through a MAS model using MATLAB software.