Multi-agent Model Research Articles

Fuzzy Multi-Agent Simulation for Collective Energy Management of Autonomous Industrial Vehicle Fleets

This paper presents a multi-agent simulation implemented in Python, using fuzzy logic to explore collective battery recharge management for autonomous industrial vehicles (AIVs) in an airport environment. This approach offers adaptability and resilience through a distributed system, taking into account variations in AIV battery capacity. Simulation scenarios were based on a proposed charging/discharging model for an AIV battery. The results highlight the effectiveness of adaptive fuzzy multi-agent models in optimizing charging strategies, improving operational efficiency, and reducing energy consumption. Dynamic factors such as workload variations and AIV-infrastructure communication are taken into account in the form of heuristics, underlining the importance of flexible and collaborative approaches in autonomous systems. In particular, an infrastructure capable of optimizing charging according to energy tariffs can significantly reduce consumption during peak hours, highlighting the importance of such strategies in dynamic environments. An optimal control model is established to improve the energy consumption of each AIV during its mission. The energy consumption depends on the speed, as demonstrated via numerical simulations using realistic data. The speed profile of each AIV is adjusted according to the various constraints within an airport. Overall, the study highlights the potential of incorporating adaptive fuzzy multi-agent models for AIV energy management to boost efficiency and sustainability in industrial operations.

Retraction Note: Visual analysis framework for network abnormal data based on multi-agent model

Retraction Note: Visual analysis framework for network abnormal data based on multi-agent model

A simulation model of negotiations on affordable housing provision

ABSTRACT In several planning systems, the provision of affordable housing is negotiated between local authorities and private developers. Those negotiations are not often accessible to students and researchers for understanding and analysis. This article presents a multi-agent simulation model of the negotiations on affordable housing in England. The model is designed to help students and researchers generate hypotheses about the impact of housing markets, local policies and negotiation behaviors on the negotiation outcomes, and consequently, the provision of affordable housing. We conducted a thought experiment using the model. To demonstrate reliability, we compared the outputs against overall trends of affordable housing provisions in England and a previous empirical study of simple negotiations. The results helped in generating four hypotheses. Among them, interestingly, the model hypothesized that there is a range of possible outcomes for each housing market, and negotiation skills can only cause variations within that range. The experiment demonstrated the model’s applicability as a tool for generating hypotheses which can open doors for future exploration, hence supporting housing education and research.

Integrating Evolutionary Game-Theoretical Methods and Deep Reinforcement Learning for Adaptive Strategy Optimization in User-Side Electricity Markets: A Comprehensive Review

With the rapid development of smart grids, the strategic behavior evolution in user-side electricity market transactions has become increasingly complex. To explore the dynamic evolution mechanisms in this area, this paper systematically reviews the application of evolutionary game theory in user-side electricity markets, focusing on its unique advantages in modeling multi-agent interactions and dynamic strategy optimization. While evolutionary game theory excels in explaining the formation of long-term stable strategies, it faces limitations when dealing with real-time dynamic changes and high-dimensional state spaces. Thus, this paper further investigates the integration of deep reinforcement learning, particularly the deep Q-learning network (DQN), with evolutionary game theory, aiming to enhance its adaptability in electricity market applications. The introduction of the DQN enables market participants to perform adaptive strategy optimization in rapidly changing environments, thereby more effectively responding to supply–demand fluctuations in electricity markets. Through simulations based on a multi-agent model, this study reveals the dynamic characteristics of strategy evolution under different market conditions, highlighting the changing interaction patterns among participants in complex market environments. In summary, this comprehensive review not only demonstrates the broad applicability of evolutionary game theory in user-side electricity markets but also extends its potential in real-time decision making through the integration of modern algorithms, providing new theoretical foundations and practical insights for future market optimization and policy formulation.

Modeling and simulation of technological innovation diffusion of high-performance construction materials in transportation engineering

This study focuses on the diffusion process of technological innovations in high-performance construction materials (HPCMs) within the field of transportation engineering. The aim is to explore the impact of various factors on this process through modeling and simulation. The research employs a multi-agent modeling approach to construct a simulation model of the diffusion of technological innovations of HPCMs among firms and owners. It analyzes the impact of firms’ strategy choices, knowledge sharing, and government policies on the diffusion of technological innovations of HPCMs. Through simulation experiments, this study reveals the impact of the proportion of firms choosing followership or innovation strategies on the efficiency of technology diffusion within complex networks, as well as the double-edged effect of knowledge sharing. Additionally, the findings indicate that adjustments of government policies at different stages of diffusion have a significant impact on market dynamics. This study provides theoretical foundations and practical recommendations for promoting the research and application of high-performance construction materials, thereby advancing green and sustainable development in transportation engineering.

Optimization of office process task allocation based on deep reinforcement learning

In office platforms, we often need to face the situation of a large number of parallel heterogeneous process tasks. This not only tests the ability of the task executors themselves, but also puts forward requirements for the performance of the collaborative scheduling system. Using the reinforcement learning method, combined with quantitative analysis such as collaborative coordination and slackness, and based on the Markov Bollinger theory, a multi-agent Bollinger model is proposed to realize the optimization scheduling system with overall process coordination and maximum completion time as the optimization objectives, thereby improving the overall execution efficiency. Taking the real business system process as the test scenario, under the same optimization objective, the reinforcement learning algorithm based on D3QN and DRL and the meta-heuristic algorithm based on ant colony are compared to verify the effectiveness of the proposed method.

Simulation of Epidemic Dynamics Using a Multi-Agent Model: Analysis of Social Distancing Strategies and Their Impacts on Public Health and Economy

Infectious disease epidemics have played a crucial role in shaping public health responses, particularly in global health crises. This study emerges as part of the efforts to prepare effective responses to potential future pandemics, leveraging lessons learned during the COVID-19 crisis. The research uses an adapted compartmental epidemiological model and a synthetic multi-agent community to investigate how social variables influence epidemic forecasts in socioeconomically vulnerable regions. Focusing on the simulation of epidemic dynamics in the socio-economically disadvantaged neighbourhood of Ilha Joana Bezerra in Recife, this study examines the impacts of social distancing strategies and other control measures, such as face masks and moderate social isolation. Through the adapted SEPAI3R3O model, which includes compartments for pre-symptomatic and asymptomatic states, this study provides a detailed analysis of disease dynamics in contexts characterised by high social vulnerability. The results underscore the importance of public health policies adapted to socio-economic factors, emphasising the need for continuous preparedness to manage future epidemic threats in vulnerable communities effectively.

Robust dynamic contracts with multiple agents

We develop a continuous-time dynamic multi-agent contracting model in which the principal is unsure about the distributions of the project's terminal payoffs and worries about model misspecification. With model uncertainty, workers' wages depend on the outputs of other unrelated projects and the optimal contracts exhibit overdetermination. We demonstrate an inverse U-shaped relationship between the extent of overdetermination and group size. Moreover, model uncertainty induces wage compression, especially in small firms as the empirical evidence demonstrates. Finally, expanding the group size increases the average project value by mitigating the negative impacts of ambiguity.

A partitioning Monte Carlo approach for consensus tasks in crowdsourcing

The planner’s role in crowdsourcing involves determining the time to stop collecting information (i.e., timed decision, TD). Previous studies have modeled the uncertain crowdsourcing environment as Partially Observable Markov Decision Processes (POMDPs) and utilized the value of information (VOI) to balance the utilities and costs of information collection. However, there is still room for optimization in solving the TD problem within single-agent POMDPs. In this paper, we propose a partitioning Monte Carlo approach for consensus tasks based on the option-candidate (OC) model partitioning. We simplify the state representation for the OC model and introduce a multi-agent POMDP representation. We establish a correspondence between the single-agent and multi-agent models using two consistency theorems. We propose three progressively improved partitioning Monte Carlo (PMC) algorithms to solve the TD problem within the multi-agent POMDP. We conducted experiments in synthetic domains and a citizen science project, demonstrating that the proposed algorithms exhibit continuous improvements in runtime advantages and consistently outperform the SOTA single-agent Monte Carlo sampling-based algorithm while providing nearly consistent output results.

HoopTransformer: Advancing NBA Offensive Play Recognition with Self-Supervised Learning from Player Trajectories.

Understanding and recognizing basketball offensive set plays, which involve intricate interactions between players, have always been regarded as challenging tasks for untrained humans, not to mention machines. In this study, our objective is to propose an artificial intelligence model that can automatically recognize offensive plays using a novel self-supervised learning approach. The dataset was collected by SportVU from 632 games during the 2015-2016 season of the National Basketball Association (NBA), with a total of 90,524 possessions. A multi-agent motion prediction pretraining model was built on the basis of axial-attention transformer and trained with different masking strategies: motion prediction (MP), motion reconstruction (MR), and MP+MR joint strategy. A downstream play-level classification task and similarity search were used to evaluate the models' performance. The results showed that the MP+MR joint masking strategy maximized the ability of the model compared with individual masking strategies. For the classification task, the joint strategy achieved a top-1 accuracy of 81.5% and top-3 accuracy of 97.5%. In the similarity search evaluation, the joint strategy attained a top-5 accuracy of 76% and top-10 accuracy of 59%. Additionally, with the same MP+MR joint masking strategy, our HoopTransformer model outperformed the two baseline models in the classification task and similarity search. This study presents a self-supervised learning model and demonstrates the effectiveness and potential of the model in accurately comprehending and capturing player movements and complex interactions during offensive plays.

Personalized Education via Mobile Technology Flexible Learning Frameworks

The prevalence of mobile and ubiquitous computing equipment is transforming how learners conduct their education. Nevertheless, the bulk of educational content generated for desktop platforms is unacceptable for handheld gadgets. Furthermore, some materials that are irrelevant to the preferences of the learner or the contextual environment might reduce the effect of learning and elevate the cost of communication. Furthermore, there is a deficit in the academic structure for mobile learning. This paper proposes personalized adaptation techniques for customized modification contents. This in turn strives to offer adaptive content contingent on learner experience and device capabilities. On top of that, it recommends a few alternatives to producing intelligent and adaptive material for children. Numerous facets of the study are examined to employ the benefits of learner, device, and location modeling in tandem with mobile technology to provide a personalized spread of multimedia-rich resources at any time or anywhere. As an outcome, it delivers proper learning resources and applies the learning style most suitable to the demands and tastes of each learner (personalized learning). Our system enables several types of learning styles to pick from, but the most significant aspect is to pair each learner’s characteristics with the right program and learning style. A novel chance for a broad range of learners to boost their knowledge in a tailored manner is made accessible by the adaptation of the material and styles, which enables the learning process to be dynamic. Additionally, our system’s resilience and interaction have risen owing to multi-agent system modeling.

The BRIDGE project: a feasibility randomised controlled trial of brief, intensive assessment and integrated formulation for young people (age 14-24) with features of borderline personality disorder (Protocol).

Borderline personality disorder (BPD) is a severe mental disorder that is characterised by a pervasive pattern of instability of interpersonal relationships, self-image, and mood as well as marked impulsivity. BPD has its peak incidence and prevalence from puberty through to emerging adulthood. BPD is a controversial diagnosis in young people. Commonly, young people with BPD are under-diagnosed, untreated, are not in employment or training and are estranged from their families. Yet, they have complex needs and are at high suicide risk. To assess the feasibility of conducting a randomised controlled trial (f-RCT) of a BRIDGE, a brief intervention programme for young people (age 14-24) with BPD symptoms (sub-threshold or threshold) in a community sample from Scotland, UK. Forty young people (age 14-24) meeting criteria for BPD symptoms, will be randomised in a 1:1 ratio to receive either a) the BRIDGE intervention plus service-as-usual or b) service-as-usual alone. Follow up will be 12 weeks and 24 weeks post-intervention. The study is carried out between 2021 to 2024. The two primary outcomes of the MQ funded, BRIDGE project (f-RCT) are i. recruitment rates and ii. retention rates. The study will present the acceptability and appropriateness of our putative outcome measures for a future definitive randomised controlled trial (d-RCT). Young people with BPD benefit from good clinical care and targeted intervention, however are regularly missed or mislabelled. The community based feasibility trial would provide initial evidence of variable needs of young people with complex needs, who maybe missed or excluded from services as they don't "fit" a model/diagnosis. Workable multi-agency service model proposed in the trial would be a major advance in understanding care pathways regardless of trial outcome. ClinicalTrial.gov, identifier NCT05023447.

Exploring the Sustainability-Oriented Strategies of Small- and Medium-Sized Construction Enterprises in China’s Construction Industry under Financing Guarantee Constraints: A Multi-Agent Computational Model Approach

In the context of China’s swift urbanization and heightened sustainability challenges, SMCEs in the construction industry play a critical role in driving both economic growth and sustainable development. This pioneering study introduces an integrative approach blending sustainability-oriented strategies (SSs) with financing guarantees, addressing key obstacles faced by these firms. Leveraging a multi-agent computational model, we dissect the dynamic interplay among construction SMCEs, governmental bodies, financial institutions, and guarantors, underscoring the transformative impact of financing guarantees on sustainable advancement. Our investigation reveals that optimally structured financing guarantees not only resolve funding challenges but also actively foster adherence to sustainability practices. This dual functionality positions such financial instruments as key enablers of SMCEs’ quest for sustainability, aligning economic objectives with ecological imperatives. Furthermore, we spotlight the transformative effect of strategic management on elevating SMCEs’ commitment to sustainable operations. Collectively, these insights offer actionable pathways for SMCEs embarking on sustainability endeavors, showcasing innovative financial solutions that do not detract from their economic prosperity.

Improvement and Analysis of Peak Shift Demand Response Scenarios of Industrial Consumers Using an Electricity Market Model

AbstractElectricity procurement of industrial consumers is becoming more and more complicated, involving a combination of various procurement methods due to electricity liberalization and decarbonization trends. This study analyzed and improved power procurement strategies for a factory to achieve carbon neutralization through a multi-agent model simulating the electricity market and introduced a factory agent using various procurement methods including PV, FC, storage batteries (SB), and DR. Firstly, we created a new procurement strategy utilizing all methods. Then, by using the simulation model, we assessed the effectiveness of the existing peak shift DR scenarios in terms of cost reduction efficiency. Results revealed that the introduction of PV has led to a counterproductive outcome for DR. Based on the results, we proposed two methods to improve the effectiveness of DR, namely considering the operation of PV in the DR scenario and expanding the range of optional time periods for DR activation. Finally, we made three new DR scenarios based on our proposal. Through experiment, the new scenarios were confirmed to be effective in cost-effectiveness for decarbonization.

Evolution of policy-driven ecosystem of original innovation talents

In an era of unprecedented global challenges, the cultivation, attraction, and retention of original innovation talent have become critical determinants of national competitiveness and sustainable development. This study introduces a novel multi-agent original innovation talent ecosystem model (MOITEM) to systematically analyze the effects of different talent policies on the ecosystems of original innovation talent across various regions in China. The simulations reveal that while direct talent policies drive short-term talent growth, they often fall short of addressing the long-term needs for resource allocation and environmental optimization, thereby limiting sustained competitiveness. In contrast, indirect and combined talent policies prioritize the overall quality of the ecosystem, providing more enduring support for original innovation. Furthermore, this study explores the interactions among agents within the ecosystem and their synergistic relationships with the external environment, highlighting the necessity of targeted policy formulation and multi-agent collaboration. The findings not only elucidate the strengths and limitations of various talent policies but also offer tailored recommendations based on regional characteristics, emphasizing the importance of continuous evaluation and adaptive policy adjustment. These insights hold substantial theoretical and practical implications for the development and optimization of global policies aimed at fostering original innovation talent.

Distributional equity impacts of automated vehicles: A disaggregated approach

Better understanding of the equity impacts of automated vehicles (AV) is needed to design equitable deployment plans for AV technology. The goal of this study was to develop and demonstrate a modeling framework to support distributional equity assessments of AV systems prior to their wide adoption. This modeling framework takes a disaggregated approach that integrates agent-based simulation, diverse transportation outcomes, and equity analyses. The framework uses individual-level data to capture detailed disparities in transportation outcomes, applies a state-of-the-art multi-agent traffic simulation model (MATSim) to simulate privately-owned and shared AVs simultaneously, and considers distributional equity from multiple perspectives. To test and demonstrate the framework, a case study was conducted for the Tampa Bay region. Five scenarios were considered with different AV market shares and integration strategies based on scenario planning by the U.S. Federal Highway Administration. Results reveal that high AV penetration rates were required for substantial reductions in inequality. The introduction of AVs led to a more even distribution of accessibility, but slightly more uneven distribution of traditional mobility and affordability. Impacts of disparities in outcomes for disadvantaged groups were mixed. These results suggest that AVs will likely perpetuate existing inequity in the transportation system as long as its fundamental structure remains the same as today. Results highlight the importance of planning and design strategies that directly address the distributional impacts to ensure that AV technology is deployed equitably.

Multi Agent Model Analysis in Identifying the Relationship Between Productivity Level, Production Cost and Product Selling Price

Multi Agent Model Analysis in Identifying the Relationship Between Productivity Level, Production Cost and Product Selling Price

A multi-agent motion simulation method for emergency scenario deduction

Simulating crowd motion in emergency scenarios remains a challenge in computer graphics due to crowd heterogeneity and environmental complexity. However, existing crowd simulation methods homogenize the agent model and simplify target selection and motion navigation of emergency crowds. To address these problems, we propose a multi-agent motion simulation method for emergency scenario deduction. First, we propose a multi-agent model to simulate crowd heterogeneity. This model includes a personality-based heterogeneous agent model and an agent perception model that considers vision, hearing, and familiarity with the environment. Second, we propose a target selection strategy based on the motion patterns of actual pedestrians. This strategy employs mathematical models and our agent perception model to guide agents in selecting appropriate targets. Finally, we propose a global navigation algorithm that combines random sampling with heuristic search methods. Concurrently, we use our multi-agent model to adjust the agent’s local motion planning to deduce the motion states of emergency crowds naturally. Experimental results validate that our method can realistically and reasonably simulate crowd motion in emergency scenarios.

Visual Reasoning and Multi-Agent Approach in Multimodal Large Language Models (MLLMs): Solving TSP and mTSP Combinatorial Challenges

Multimodal Large Language Models (MLLMs) harness comprehensive knowledge spanning text, images, and audio to adeptly tackle complex problems. This study explores the ability of MLLMs in visually solving the Traveling Salesman Problem (TSP) and Multiple Traveling Salesman Problem (mTSP) using images that portray point distributions on a two-dimensional plane. We introduce a novel approach employing multiple specialized agents within the MLLM framework, each dedicated to optimizing solutions for these combinatorial challenges. We benchmarked our multi-agent model solutions against the Google OR tools, which served as the baseline for comparison. The results demonstrated that both multi-agent models—Multi-Agent 1, which includes the initializer, critic, and scorer agents, and Multi-Agent 2, which comprises only the initializer and critic agents—significantly improved the solution quality for TSP and mTSP problems. Multi-Agent 1 excelled in environments requiring detailed route refinement and evaluation, providing a robust framework for sophisticated optimizations. In contrast, Multi-Agent 2, focusing on iterative refinements by the initializer and critic, proved effective for rapid decision-making scenarios. These experiments yield promising outcomes, showcasing the robust visual reasoning capabilities of MLLMs in addressing diverse combinatorial problems. The findings underscore the potential of MLLMs as powerful tools in computational optimization, offering insights that could inspire further advancements in this promising field.

A multi-agent curiosity reward model for task-oriented dialogue systems

In practical decision-making dialogues, reinforcement learning methods face hurdles due to delays and sparse reward feedback for agents, and in some cases, lack of rewards altogether. These issues can impede efficient learning of dialogue strategies and compromise the performance of the model. To address this challenge, this paper introduces the Multi-Agent Curiosity Reward Model (MACRM) for task-oriented dialog systems. Firstly, in terms of dialog reward mechanisms, a forward dynamics model generates curiosity rewards, which are integrated with extrinsic rewards from the dialog environment feedback to mitigate the problem of sparse rewards resulting from inadequate agent exploration. Secondly, regarding the dialogue strategy training mechanism, an exploration-exploitation approach inspired by organismic exploration is adopted. This approach involves fully exploring the dialogue environment in the early stages and optimally exploiting learned knowledge later, thereby balancing exploration and exploitation and enhancing dialogue strategy learning efficiency. To assess the proposed model's effectiveness, experiments are conducted using the MultiWOZ corpus across three reward environments: (1) extrinsic rewards only, (2) curiosity rewards only, and (3) a combination of both. The experimental results demonstrate that agents employing MACRM exhibit faster learning of dialogue strategies compared to those relying on a single exploratory reward method, effectively addressing reward sparsity and delay issues in practical decision-making scenarios.