Agent-based System Research Articles

Parameter Tuning of Agent-Based Models: Metaheuristic Algorithms

When it comes to modelling complex systems using an agent-based approach, there is a problem of choosing the appropriate parameter optimisation technique. This problem is further aggravated by the fact that the parameter space in complex agent-based systems can have a large dimension, and the time required to perform numerical experiments can be large. An alternative approach to traditional optimisation methods are the so-called metaheuristic algorithms, which provide an approximate solution in an acceptable time. The purpose of this study is to compare various metaheuristic algorithms for parameter tuning and to analyse their effectiveness applied to two agent-based models with different complexities. In this study, we considered commonly used metaheuristic algorithms for agent-based model optimisation: the Markov chain Monte Carlo method, the surrogate modelling approach, the particle swarm optimisation algorithm, and the genetic algorithm, as well as the more novel chaos game optimisation algorithm. The proposed algorithms were tested on two agent-based models, one of which was a simple toy model of the spread of contagious disease, and the other was a more complex model of the circulation of respiratory viruses in a city with 10 million agents and 26 calibrated parameters.

Machine learning for the identification of phase transitions in interacting agent-based systems: A Desai-Zwanzig example.

Deriving closed-form analytical expressions for reduced-order models, and judiciously choosing the closures leading to them, has long been the strategy of choice for studying phase- and noise-induced transitions for agent-based models (ABMs). In this paper, we propose a data-driven framework that pinpoints phase transitions for an ABM-the Desai-Zwanzig model-in its mean-field limit, using a smaller number of variables than traditional closed-form models. To this end, we use the manifold learning algorithm Diffusion Maps to identify a parsimonious set of data-driven latent variables, and we show that they are in one-to-one correspondence with the expected theoretical order parameter of the ABM. We then utilize a deep learning framework to obtain a conformal reparametrization of the data-driven coordinates that facilitates, in our example, the identification of a single parameter-dependent ordinary differential equation(ODE) in these coordinates. We identify this ODE through a residual neural network inspired by a numerical integration scheme (forward Euler). We then use the identified ODE-enabled through an odd symmetry transformation-to construct the bifurcation diagram exhibiting the phase transition.

Free Riding of Vehicle Companies under Dual-Credit Policy: An Agent-Based System Dynamics Model

The dual-credit policy promotes green transition in automobile companies. This paper investigates the dual-credit policy framework in the Chinese automotive industry, with a focus on the phenomenon of free riding. This occurs when traditional vehicle manufacturers within an alliance benefit from the excess credits generated by a transitioning vehicle company without fully committing to their own green transitioning. The focus of this study lies on an alliance constituted by a transitioning vehicle company in partnership with two traditional vehicle manufacturers, all interconnected via equity ties. Utilizing an agent-based system dynamics model, this study explores the strategic behaviors emerging from such credit collaborations and their consequent effects on operational efficiency and financial performance. The findings reveal that 1. free riding negatively impacts the transitioning company’s revenue but benefits the alliance by easing transition pressures and boosting collective performance; 2. stricter policies increase intra-alliance credit transfers and performance, while lower credit prices reduce transfer value and harm the transitioning company’s earnings. This study implies that transitioning vehicle companies with equity-linked partners can benefit from a nuanced understanding of how policy mechanisms interact with alliance dynamics under free riding. By adjusting credit transfer strategies in line with market conditions and policy trends, they can better navigate the dual-credit policy landscape, balancing individual profitability with the needs of the broader alliance and long-term sustainability goals.

Forecasting municipal solid plastic waste generation and management policy using system dynamics: a case study of Khulna City in Bangladesh

A comprehensive analysis of municipal solid plastic waste (MSPW) management while emphasizing plastic pollution severity in coastal cities around the world is mandatory to alleviate the augmenting plastic waste footprint in nature. Thus, decision-makers’ persuasion for numerous management solutions of MSPW flow-control can be met through meditative systematic strategies at the regional level. To forecast solutions focused on systematic policies, an agent-based system dynamics (ASD) model has been developed and simulated from 2023 to 2040 while considering significant knit parameters for MSPW management of Khulna City in Bangladesh. Baseline simulation results show that per-capita plastic waste generation will increase to 11.6 kg by 2040 from 8.92 kg in 2023. Eventually, the landfilled quantity of plastic waste has accumulated to 70,000 tons within 18 years. Moreover, the riverine discharge has increased to 834 tons in 2040 from a baseline quantity of 512 tons in 2023. So the plastic waste footprint index (PWFI) value rises to 24 by 2040. Furthermore, the absence of technological initiatives is responsible for the logarithmic rise of non-recyclable plastic waste to 1.35*1000=1350 tons. Finally, two consecutive policy scenarios with baseline factors such as controlled riverine discharge, increased collection and separation of plastic waste, expansion of recycle business, and locally achievable plastic conversion technologies have been simulated. Therefore, policy 2, with 69% conversion, 80% source separation, and 50% riverine discharge reduction of MSPW, has been found adequate from a sustainability perspective with the lowest PWFI ranges of 3.97 to 1.07 alongside a per-capita MSPW generation of 7.63 to 10 kg from 2023 till 2040.Graphical abstract

Agent-Based System for Continuous Control and its Application to Activated Sludge Process

This paper presents a concept of architecture and ontology layouts for development of multiagent model-based predictive control systems. The presented architecture provides guidelines to simplify development of agent-based systems and improve its maintainability. The proposed multiagent system (MAS) layout is split into multiple subsystems that include agents dedicated to performing assigned tasks. MAS implementation was prepared which can use provided algorithms, actuators and can react to changes in its environment to reach the best available control quality. An example of MAS based on the proposed architecture is shown in application of dissolved oxygen (DO) concentration control in a laboratory activated sludge setup with biological reactor. For that application, MAS incorporates agent-based controllers from Boundary-Based Predictive Controllers (BBPC) family. Presented experiments prove flexibility, resilience and online reconfiguration ability of the proposed multiagent system.

AI Concepts for System of Systems Dynamic Interoperability.

Interoperability is a central problem in digitization and System of Systems (SoS) engineering, which concerns the capacity of systems to exchange information and cooperate. The task to dynamically establish interoperability between heterogeneous cyber-physical systems (CPSs) at run-time is a challenging problem. Different aspects of the interoperability problem have been studied in fields such as SoS, neural translation, and agent-based systems, but there are no unifying solutions beyond domain-specific standardization efforts. The problem is complicated by the uncertain and variable relations between physical processes and human-centric symbols, which result from, e.g., latent physical degrees of freedom, maintenance, re-configurations, and software updates. Therefore, we surveyed the literature for concepts and methods needed to automatically establish SoSs with purposeful CPS communication, focusing on machine learning and connecting approaches that are not integrated in the present literature. Here, we summarize recent developments relevant to the dynamic interoperability problem, such as representation learning for ontology alignment and inference on heterogeneous linked data; neural networks for transcoding of text and code; concept learning-based reasoning; and emergent communication. We find that there has been a recent interest in deep learning approaches to establishing communication under different assumptions about the environment, language, and nature of the communicating entities. Furthermore, we present examples of architectures and discuss open problems associated with artificial intelligence (AI)-enabled solutions in relation to SoS interoperability requirements. Although these developments open new avenues for research, there are still no examples that bridge the concepts necessary to establish dynamic interoperability in complex SoSs, and realistic testbeds are needed.

A Systematic Approach to Autonomous Agents

Agents and agent-based systems are becoming essential in the development of various fields, such as artificial intelligence, ubiquitous computing, ambient intelligence, autonomous computing, and intelligent robotics. The concept of autonomous agents, inspired by the observed agency in living systems, is also central to current theories on the origin, development, and evolution of life. Therefore, it is crucial to develop an accurate understanding of agents and the concept of agency. This paper begins by discussing the role of agency in natural systems as an inspiration and motivation for agential technologies and then introduces the idea of artificial agents. A systematic approach is presented for the classification of artificial agents. This classification aids in understanding the existing state of the artificial agents and projects their potential future roles in addressing specific types of problems with dedicated agent types.

Koopman-based surrogate models for multi-objective optimization of agent-based systems

Agent-based models (ABMs) provide an intuitive and powerful framework for studying social dynamics by modeling the interactions of individuals from the perspective of each individual. In addition to simulating and forecasting the dynamics of ABMs, the demand to solve optimization problems to support, for example, decision-making processes naturally arises. Most ABMs, however, are non-deterministic, high-dimensional dynamical systems, so objectives defined in terms of their behavior are computationally expensive. In particular, if the number of agents is large, evaluating the objective functions often becomes prohibitively time-consuming. We consider data-driven reduced models based on the Koopman generator to enable the efficient solution of multi-objective optimization problems involving ABMs. In a first step, we show how to obtain data-driven reduced models of non-deterministic dynamical systems (such as ABMs) that depend potentially nonlinearly on control inputs. We then use them in the second step as surrogate models to solve multi-objective optimal control problems. We first illustrate our approach using the example of a voter model, where we compute optimal controls to steer the agents to a predetermined majority, and then using the example of an epidemic ABM, where we compute optimal containment strategies in a prototypical situation. We demonstrate that the surrogate models effectively approximate the Pareto-optimal points of the ABM dynamics by comparing the surrogate-based results with test points, where the objectives are evaluated using the ABM. Our results show that when objectives are defined by the dynamic behavior of ABMs, data-driven surrogate models support or even enable the solution of multi-objective optimization problems.

A Multiple Agent-Based System for the Intelligent Demand Planning of New Products

A Multiple Agent-Based System for the Intelligent Demand Planning of New Products

A multiple agent-based system for the intelligent demand planning of new products

A multiple agent-based system for the intelligent demand planning of new products

Effects of a Gamified Agent-Based System for Personalized Elderly Care: Pilot Usability Study.

The global percentage of older people has increased significantly over the last decades. Information and communication technologies have become essential to develop and motivate them to pursue healthier ways of living. This paper examines a personalized coaching health care service designed to maintain living conditions and active aging among older people. Among the technologies the service includes, we highlight the use of both gamification and cognitive assistant technologies designed to support older people and an application combining a cognitive virtual assistant to directly interact with the older person and provide feedback on their current health condition and several gamification techniques to motivate the older person to stay engaged with the application and pursuit of healthier daily habits. This pilot study aimed to investigate the feasibility and usability of a gamified agent-based system for older people and obtain preliminary results on the effectiveness of the intervention regarding physical activity health outcomes. The study was designed as an intervention study comparing pre- and posttest results. The proposed gamified agent-based system was used by 12 participants over 7 days (1 week), and step count data were collected with access to the Google Fit application programming interface. Step count data after the intervention were compared with average step count data before the intervention (average daily values over a period of 4 weeks before the intervention). A 1-tailed Student t test was used to determine the relationship between the dependent and independent variables. Usability was measured using the System Usability Scale questionnaire, which was answered by 8 of the 12 participants in the study. The posttest results showed significant pre- to posttest changes (P=.30; 1-tailed Student t test) with a moderate effect size (Cohen d=0.65). The application obtained an average usability score of 78. The presented pilot was validated, showing the positive health effects of using gamification techniques and a virtual cognitive assistant. Additionally, usability metrics considered for this study confirmed high adherence and interest from most participants in the pilot.

Multi-Agent Systems Requirements Analysis for Patient-centered Healthcare Consultancy Service

Agents are currently being discussed in nearly every domain of science and engineering. Through the utilization of IoT and agent-based systems, remote consultations and virtual doctors can provide essential healthcare services by analyzing patient data and medical history. The purpose of this paper is provide an analysis the requirements for multi-agent systems in patient-centered health consultancy services in Kenya. It focuses on various agent types, including hardware, hospital, human, software, database/storage, and technological/framework agents. For hardware, computers, laptops, and mobile devices need specific processing power, memory, network connectivity, storage, and security. Medical equipment must prioritize accuracy, reliability, integration with records, and regulatory compliance. Networking requires high-speed internet, security, and redundancy. Hospital agents offer comprehensive inpatient and outpatient care, categorized into general, mental health, and specialized hospitals. Human agents, like medical specialists, IT professionals, and administrators, play essential roles in healthcare consultancy. Software agents encompass Electronic Health Records (EHRs), Clinical Decision Support Systems (CDSS), Telemedicine platforms, and more. Database/storage agents include EHRs, Health Information Exchange (HIE) systems, and data encryption. Technological/framework agents involve health information systems, telemedicine platforms, decision support systems, and health data analytics frameworks. These agents enable efficient data management, remote care, evidence-based decisions, and secure data exchange.

Simulation Modelling for the Promotion of Green Residence Based on the Theory of Sustainability—Taking Jiangsu Province as an Example

Green residences have enormous potential for energy savings, emission reduction, and other comprehensive benefits, and their growth is crucial to achieving China’s carbon neutrality and carbon peaking targets. Nevertheless, at the moment, the national green residence is impacted by complicated factors at several levels, including government agencies, green residence builders, and green residence consumers, which results in the low-quality development of domestic green residences overall. As of 2020, 94% of all labeled green residences are design-label residences that can only be achieved during the design stage, while less than 10% are operational-label residences with stronger energy and emission-saving benefits. This causes the phenomenon of “green residences on the planning” to be serious. In order to accomplish the promotion of high-quality development of green residences and to promote green residences in China, this paper analyzes the influencing factors of green residence promotion from the multi-level perspective of macro-landscape signals, meso-collective agent green residences, and micro-individual agent consumers, based on the multi-level perspective (MLP) framework of sustainability theory. The paper subsequently builds a simulation model of green residence promotion using the agent-based system dynamics modeling method. Additionally, Jiangsu Province’s green residence promotion data are chosen for analogue simulation experiments, and the simulation results are also used to analyze the success conditions as well as the path to green residence promotion. This study demonstrates that (1) the agent-based simulation model of dynamics for the green residence promotion system has high reference value for the simulation of the promotion of green residences, and the model can clearly simulate the impact of micro-individual agent–consumer factors on the promotion of green residences; (2) in order to promote green residences, exterior landscape signals must be continuously improved; the stronger the landscape signals, the quicker the development of operationally labeled green residences; (3) priority is given to the development of two-star design-labeled green residences before 2035, and three-star operationally labeled residences will occupy the majority of the market after 2040. Meanwhile, the duration of landscape signals and the change in behavioral preferences of individual agents must be maintained for a long time.

Knowledge Management in Multimedia Communication Using Software Agents

This paper presents the problem of knowledge management related to software agents and agent systems (i.e. systems composed of communicating and cooperating software agents). The author presents basic theoretical issues, but the main focus is on the areas of application of software agents in multimedia communication - both existing and prospective applications. Particular attention has been paid to the applications of software agents for: supporting customer service processes through the Contact Centre system; supporting the distance learning system; creating virtual assistants in the form of conversation bots; searching for information in the Semantic Web, and assisting in setting up a video conference call. Software agents are often referred to as 'intelligent' - due to their intelligent behaviour. At their core, they involve: perception, reasoning, learning, communication and operation in complex environments, using all multimedia techniques and procedures. Software agents and agent systems, as presented in this article, represent a new approach to the analysis, design and implementation of complex (usually decentralised and distributed) computer systems, offering a whole range of concepts, techniques and tools to significantly improve the software development process. The ability of software agents to make autonomous decisions have been a subject of considerations. Among other things, they are based on the example of cognitive sciences, which deal with modelling real processes of information processing in humans and building programmes that imitate these processes. Research on the common understanding of shared knowledge and knowledge management in agent-based systems focuses on ontologies, understood as a description of the relationship between the way knowledge is represented and the concept represented by that knowledge. In contrast, research on knowledge sharing has focused on communication languages and their associated - interaction protocol and message transport protocol.

Telemedicine, E-Health, and Multi-Agent Systems for Chronic Pain Management.

Telemedicine, telehealth, and E-health all offer significant benefits for pain management and healthcare services by fostering the physician-patient relationship in otherwise challenging circumstances. A critical component of these artificial-intelligence-based health systems is the "agent-based system", which is rapidly evolving as a means of resolving complicated or straightforward problems. Multi-Agent Systems (MAS) are well-established modeling and problem-solving modalities that model and solve real-world problems. MAS's core concept is to foster communication and cooperation among agents, which are broadly considered intelligent autonomous factors, to address diverse challenges. MAS are used in various telecommunications applications, including the internet, robotics, healthcare, and medicine. Furthermore, MAS and information technology are utilized to enhance patient-centered palliative care. While telemedicine, E-health, and MAS all play critical roles in managing chronic pain, the published research on their use in treating chronic pain is currently limited. This paper discusses why telemedicine, E-health, and MAS are the most critical novel technologies for providing healthcare and managing chronic pain. This review also provides context for identifying the advantages and disadvantages of each application's features, which may serve as a useful tool for researchers.

Horizontal collaboration between suppliers to mitigate supply chain disruption: A secure resource sharing strategy

Resource sharing is a flexible recovery strategy that has attracted attention with the outbreak of global disruption. Horizontal cooperation could lead to mutual benefits and reduce the negative consequences during the recovery period. This paper emphasizes the advantages of sharing resources between two competing suppliers under disruption. Suppliers share a secure source of inventory to absorb the shocks of disruption. The stability of relationships between suppliers is sensitive to several factors. Accordingly, an evolutionary game model is developed followed by a replicator dynamic analysis to study the stability of strategies under disruption scenarios. A hybrid agent-based system dynamics framework is built using AnyLogic software to simulate the relationships, interactions, and the effect of parameters in the adopted strategies of suppliers. Compared with mutual incentives and penalty costs, findings indicate that the value of secure resources and the number of shortages have higher impacts on the stability of cooperation. A higher rate of shortages could undermine collaboration. Results show that the initial mindset of suppliers to follow a collaborative approach is a key driver for the stable implementation of collaborative resource sharing. This paper highlights the benefits of implementing new technologies to improve supply flexibility through horizontal integration with competitors.

Optimized Load Frequency Controller Performances With a Reduced Number of Decision Variables: A Gravitational Search Algorithm-Based Method

Ensuring stability and maintaining performances of the load frequency controller (LFC) are stimulating tasks. A state feedback controller (SFC) with a generalized range of operation satisfying the necessary and sufficient conditions for system stability along with optimized performances through the gravitational search algorithm (GSA) are proposed. The proposed range is dependent on the system parameters. The agent-based systems in the literature consider the dimension of an agent as per the number of decision variables of the problem. The proposed method reduces the number of decision variables to obtain the controller gains and to use in the optimization process. The proposed method is validated through extensive results for different objective functions, contemplating both the steady state and the transient requirements as well as a consideration of different perturbed system parameters and variations in the input disturbances. Results and comparative studies with popular control techniques and optimization algorithms testify that the proposed method maintains system stability and improves the system performances while considering different system parameters, parametric uncertainties, and variations in input disturbances.

Using active learning and an agent-based system to perform interactive knowledge extraction based on the COVID-19 corpus

Abstract Efficient knowledge extraction from Big Data is quite a challenging topic. Recognizing relevant concepts from unannotated data while considering both context and domain knowledge is critical to implementing successful knowledge extraction. In this research, we provide a novel platform we call Active Learning Integrated with Knowledge Extraction (ALIKE) that overcomes the challenges of context awareness and concept extraction, which have impeded knowledge extraction in Big Data. We propose a method to extract related concepts from unorganized data with different contexts using multiple agents, synergy, reinforcement learning, and active learning. We test ALIKE on the datasets of the COVID-19 Open Research Dataset Challenge. The experiment result suggests that the ALIKE platform can more efficiently distinguish inherent concepts from different papers than a non-agent-based method (without active learning) and that our proposed approach has a better chance to address the challenges of knowledge extraction with heterogeneous datasets. Moreover, the techniques used in ALIKE are transferable across any domain with multidisciplinary activity.

Internet of things and agent-based system to improve water use efficiency in collective irrigation

The efficient management of water resources is a major issue in the field of sustainable development. Several models of solving this problem can be found in the literature, especially in the agricultural sector which represents the main consumer through irrigation. Therefore, Irrigation management is an important and innovative field that has been the subject of several types of research and studies to deal with the different activities, behaviors, and conflicts between the different users. This article introduces an intelligent irrigation system based on smart sensors that can be used moderately and economically to monitor farms by integrating some connected electronic devices and other advantageous instruments widely used in the field of IoT, it determines the water requirement of each farm according to the water loss due to the process of evapotranspiration. The water requirement is calculated from data collected from a series of sensors installed in the plantation farm. This project focuses on smart irrigation based on IoT and agent technology, it can be used by farmer associations whose endowments and irrigation planning are defined according to the need and quantity of water available in the rural municipality. The system includes a microcontroller with the integration of sensors, actuators, and valve modules where each node serves as an IoT device. Environmental parameters are monitored directly through a multi-agent system that facilitates the control of each node and the configuration of irrigation parameters. The amount of water calculated for irrigation is based on the Penman model for calculating the daily evapotranspiration baseline. Compared to the conventional irrigation method, it is expected that the proposed irrigation model would contribute to saving water use and distributing it impartially without compromising its production.

Mobile value chain collaboration for product diffusion: Role of the lifecycle

To succeed in a competitive market, mobile value chains (MVC) must respond to dynamic demand timely. To offer effective collaboration mechanisms in MVC for mobile service diffusion over the lifecycle, we propose a three-stage integrated simulation model: (i) a system dynamics model to depict mobile service diffusion which drives the evolution of MVC; (ii) a cooperative game model to examine the profit allocation between mobile operators and providers; (iii) an evolution game model to explore the collaboration in the mobile alliance of providers. Then we design an agent-based system to replicate the evolution process of MVC under various settings, and further validate it by comparison with a real-world case. We find that an MVC would need different strategies at different lifecycle stages. Advertising and the connection degree of mobile operators contribute significantly to MVC performance at the growth stage. In contrast, word-of-mouth (WOM) and the penalty is more effective at the mature stage; penalty can extend the mature stage. Finally, we find the scale and the components ratio of a mobile alliance positively affect MVC throughout its lifecycle and decline stage, respectively. Thus, management should select fitting mechanisms over the lifecycle to improve its performance.