Smart transportation networks are challenged by a critical challenge of traffic congestion and inevitably requires a complex package of management solutions. This paper introduces a Secure Multi-Agent Reinforcement Learning (MARL) framework that optimizes traffic whilst protecting from malicious threats. Traffic Prediction Dataset accessible through Kaggle is used as the foundation for training reinforcement learning agents from four junctions who cooperate to manage traffic signals. The widespread preprocessing procedures that also dealt with data gaps in addition to dimension transformation and value normalization generated genuine traffic prediction models. The Multi-Agent Reinforcement Learning framework reduced the mean waiting time by 45.7% with a 47.0% traffic passing rate higher than that achieved using standard fixed-duration signal control techniques. Homomorphic encryption, when integrated with blockchain protocols, instituted novel security protocols for agent to agent communication, and this yielded a reduction of adversarial attacks below 90% levels. The implementation framework provided a 39.3% reduction in the fuel consumption when implemented in real-time urban traffic management systems. Secure Multi-Agent Reinforcement Learning (MARL) systems show their potential aptitude to propel smart transportation networks in the while providing increased efficiency and security as well as sustainability effectiveness.
A Multi-Agent Reinforcement Learning Framework for Public Health Decision Analysis.
A Multi-Agent Reinforcement Learning Framework for Public Health Decision Analysis.
Hybrid Multi-Agent Reinforcement Learning and Blockchain Framework for Real-Time Transaction Integrity in Cloud-Driven Financial Systems
Financial institutions are increasingly dealing with difficulties in preserving transaction integrity due to escalating digital fraud and extensive cloud-based operations. Conventional machine learning and rule-based detection systems frequently have latency in replies, restricted adaptability, and an absence of verifiable audit trails. The proposed study presents a Hybrid Multi-Agent Reinforcement Learning (MARL) and Blockchain framework to overcome existing limitations, integrating adaptive multi-agent coordination for real-time fraud detection with a permissioned blockchain ledger to guarantee secure, tamper-proof validation of agent decisions. MARL agents adaptively acquire optimal detection strategies through incentive shaping and inter-agent communication, while blockchain consensus and smart contracts substantiate decision proofs for regulatory adherence. The experimental evaluation of the Kaggle Financial Transactions dataset demonstrates improved accuracy relative to baseline models: Accuracy 0.989, Precision 0.981, Recall 0.987, F1-Score 0.984, AUC-ROC 0.992, with negligible latency (1.8 ms/transaction). The comparative research reveals a persistent superiority over Random Forest, XGBoost, CNN- LSTM, and Transformer-based classifiers, showcasing robustness, scalability, and real-time application. The suggested study enhances the area by merging reinforcement learning adaptability with blockchain security, establishing a realistic, high-throughput, and reliable framework for cloud- based financial systems, while presenting a model for future research in agentic AI-driven fraud prevention.
Read more
4A Deep Reinforcement Learning and Graph Convolution Approach to On-Street Parking Search Navigation.
Efficient parking distribution is crucial for urban traffic management; nevertheless, variable demand and spatial disparities raise considerable obstacles. Current research emphasizes local optimization but neglects the fundamental challenges of real-time parking allocation, resulting in inefficiencies within intricate metropolitan settings. This research delineates two key issues: (1) A dynamic imbalance between supply and demand, characterized by considerable fluctuations in parking demand over time and across different locations, rendering static allocation solutions inefficient; (2) spatial resource optimization, aimed at maximizing the efficiency of limited parking spots to improve overall system performance and user satisfaction. We present a Multi-Agent Reinforcement Learning (MARL) framework that incorporates adaptive optimization and intelligent collaboration for dynamic parking allocation to tackle these difficulties. A reinforcement learning-driven temporal decision mechanism modifies parking assignments according to real-time data, whilst a Graph Neural Network (GNN)-based spatial model elucidates inter-parking relationships to enhance allocation efficiency. Experiments utilizing actual parking data from Melbourne illustrate that Multi-Agent Reinforcement Learning (MARL) substantially surpasses conventional methods (FIFO, SIRO) in managing demand variability and optimizing resource distribution. A thorough quantitative investigation confirms the strength and flexibility of the suggested method in various urban contexts.
Read more6Enhancing collaboration in uncertain environment: Multi-Agent Reinforcement Learning for underwater monitoring
Underwater monitoring is extremely complex due to the lack of a global localization system, limited communication and environmental factors such as turbidity and darkness that limit visibility, affecting control and situational awareness. Typically, monitoring relies on a single autonomous underwater vehicle (AUV) or a set of independent AUVs; techniques which are prone to failure as they rely only on onboard odometry and sensors, making missions vulnerable to malfunctions, damage, and noise. To address these challenges, we propose a Multi-Agent Reinforcement Learning (MARL) framework to enable cooperation among multiple AUVs, mitigating the limitations of the underwater environment. Our in-silico solution focuses on a group of robots learning a strategy to follow a partially hidden underwater pipe without global localization, while dealing with environmental disturbances affecting sensors and actuators. The numerosity of the agents, and most importantly their collaboration, helps overcome underwater visibility constraints. By sharing relative position information of neighboring agents with respect to the pipe, navigation is improved. By introducing quantitative measures for pipe exploration, we show that cooperation significantly enhances system performance compared to independent agents. Emerging collaboration among robots allows the swarm to complete pipe inspections faster and more efficiently than non-cooperative baseline models of non-interacting agents, even under extremely reduced visibility scenarios. Moreover, single agents also benefit from cooperation, learning effective policies more quickly and covering a longer portion of the pipe. Finally, our model guarantees explainability. We analyze learned strategies and provide a visualization method that allows the interpretation of the learned policies. • Reinforcement learning is applied to pipeline following by underwater robotic agent. • In conditions of poor visibility, a single agent is not able to complete the mission. • In contrast, multi-agent team completes this task using reinforcement learning. • Swarm collaboration enables faster and more efficient task completion. • Collaboration enables agents develop more efficient individual strategies.
Read more5Video Service-Oriented Vehicular Collaboration: A Multi-Agent Proximal Policy Optimization Approach
To guarantee heterogeneous performance requirements of diverse vehicular services, it is necessary to design a full cooperative policy for both vehicle to infrastructure (V2I) links and vehicle to vehicle (V2V) links. This paper investigates how to improve the quality of experience (QoE) of the V2I users for video services while satisfying the delay requirements of both V2I and V2V links. In specific, a QoE maximization problem is formulated with consideration of vehicular collaboration where task offloading decision, channel reuse decision and power allocation of V2V users are all included. A multi-agent reinforcement learning (MARL) framework is then designed, where a new reward function is proposed to evaluate the utility of the considered network. Thereafter, a proximal policy optimization approach is proposed to enable each V2V user to learn policy individually with the shared global network reward. The effectiveness of the proposed approach is finally validated with comparison of other baseline approaches through extensive simulation experiments.
Read more7Searching Collaborative Agents for Multi-plane Localization in 3D Ultrasound
3D ultrasound (US) has become prevalent due to its rich spatial and diagnostic information not contained in 2D US. Moreover, 3D US can contain multiple standard planes (SPs) in one shot. Thus, automatically localizing SPs in 3D US has the potential to improve user-independence and scanning-efficiency. However, manual SP localization in 3D US is challenging because of the low image quality, huge search space and large anatomical variability. In this work, we propose a novel multi-agent reinforcement learning (MARL) framework to simultaneously localize multiple SPs in 3D US. Our contribution is four-fold. First, our proposed method is general and it can accurately localize multiple SPs in different challenging US datasets. Second, we equip the MARL system with a recurrent neural network (RNN) based collaborative module, which can strengthen the communication among agents and learn the spatial relationship among planes effectively. Third, we explore to adopt the neural architecture search (NAS) to automatically design the network architecture of both the agents and the collaborative module. Last, we believe we are the first to realize automatic SP localization in pelvic US volumes, and note that our approach can handle both normal and abnormal uterus cases. Extensively validated on two challenging datasets of the uterus and fetal brain, our proposed method achieves the average localization accuracy of 7.03 degrees/1.59mm and 9.75 degrees/1.19mm. Experimental results show that our light-weight MARL model has higher accuracy than state-of-the-art methods.
Read moreCrowd-Aware Itinerary Optimization via Clustered Multi-Agent Reinforcement Learning
Personalized travel sequence recommendation in urban environments poses a complex challenge due to dynamic spatial conditions, user diversity and limited resources. Traditional systems often overlook the multi-user dimension and fail to adapt to fluctuating crowd patterns. This work proposes a novel Multi-Agent Reinforcement Learning (MARL) framework that coordinates personalized itinerary planning while managing urban congestion. The framework comprises autonomous Travel Agents representing individual users and a centralized Congestion Management Authority that guides agents toward balanced spatial distributions. To enhance scalability and learning efficiency we introduce an interest-based clustering mechanism that groups users with similar preferences. Each cluster shares a policy network and experience buffer, and is trained using a Deep Q-Network (DQN) algorithm. This design significantly reduces memory consumption by 80% and training time by 67%. Compared to the baseline model our MARL framework improves interest alignment by 49.1% (0.85 vs. 0.57), PoI popularity by 50.9% (0.80 vs. 0.53) and reduces travel time by 50.5% (11.16% vs. 22.55%) resulting in more relevant, attractive and time-efficient itineraries. A real-world case study in Dubai validates the framework's ability to generate high-quality itineraries that align with user interests and mitigate overcrowding at popular sites. Additionally, a user study confirms improved satisfaction and perceived personalization compared to baseline methods. The results highlight the potential of MARL as a scalable, adaptive solution for next-generation spatial recommendation systems.
Read moreFederated Multi-agent Reinforcement Learning for Edge-intelligent Beamforming Codebook Design for mmWave MIMO
Hybrid beamforming with massive multiple-input multiple-output (MIMO) antenna arrays plays a pivotal role in achieving high data rates and spectral efficiency in next-generation wireless systems. Nevertheless, the real-time 62design and adaptation of beamforming codebooks and user grouping remain challenging, as conventional machine learning (ML) methods typically rely on the centralized collection of extensive channel state information (CSI), resulting in substantial feedback overhead and potential privacy violations. This chapter proposes a federated multi-agent reinforcement learning (MARL) framework for joint hybrid beamforming and dynamic user grouping. Unlike prior work which focuses solely on centralized codebook learning, the proposed approach enables distributed user equipment (UE) to collaboratively train codebook and clustering algorithms without sharing raw data, thereby preserving privacy. The proposed method significantly reduces communication overhead while achieving near-optimal beam patterns and robust scalability in dense network scenarios, making it well-suited for practical deployment in next-generation wireless systems. The experimental findings indicate that for mobile users, federated learning (FL)-based user clustering with MARL outperforms traditional initial access-based beam clustering in MARL.
Read more4Optimal tracking agent: a new framework of reinforcement learning for multiagent systems
SUMMARYThe curse of dimensionality is a ubiquitous problem for multiagent reinforcement learning, which means the learning and storing space grows exponentially with the number of agents and hinders the application of multiagent reinforcement learning. To relieve this problem, we propose a new framework named as optimal tracking agent (OTA). The OTA views the other agents as part of the environment and uses a reduced form to learn the optimal decision. Although merging other agents into the environment may reduce the dimension of action space, the environment characterized by such form is dynamic and does not satisfy the convergence of reinforcement learning (RL). Thus, we develop an estimator to track the dynamics of the environment. The estimator obtains the dynamic model, and then the model‐based RL can be used to react to the dynamic environment optimally. Because the Q‐function in OTA is also a dynamic process because of other agents’ dynamics, different from traditional RL, in which the learning is a stationary process and the usual action selection mechanisms just suit to such stationary process, we improve the greedy action selection mechanism to adapt to such dynamics. Thus, the OTA will have convergence. An experiment illustrates the validity and efficiency of the OTA.Copyright © 2012 John Wiley & Sons, Ltd.
Read more2A Multi-Agent Reinforcement Learning Framework with Recurrent Communication Module for Traffic Light Control
As a significant issue in the development of Smart City, traffic light control is closely related to public transport. Multi-agent reinforcement learning is a feasible paradigm to control decentralized systems. In this work, we design a recurrent communication module for traffic light control systems. Aiming at lightweight decentralized learning, we decouple the interaction among entities, using an improved recurrent neural network to share messages. Regarding traffic lights as a partially observable Markov Decision Process, we develop a complete reinforcement learning model. Every light receives messages from its neighbors through the improved recurrent neural network. For the further lights, its neighbor lights can play the role of intermediation to transmit messages. By our communication framework, traffic lights can learn their control policies in a decentralized way, and distribution control can be executed through a communication channel. Several simulated experiments show that our method outperforms the fixed time method and the existing reinforcement learning-based frameworks under different traffic demands.
Read more26Multi-agent reinforcement learning for edge information sharing in vehicular networks
Multi-agent reinforcement learning for edge information sharing in vehicular networks
2Multi-Agent Reinforcement Learning Based Electric Vehicle Charging Control for Grid-Level Services
Coordinating the charging process of a large population of electric vehicles (EV) is promising in increasing power grid flexibility from the demand side, yet requires highly scalable control protocols. In contrast to classical decentralized optimization based methods that require approximated distribution network models, this paper frames the EV charging control problem into a multi-agent reinforcement learning (MARL) framework. The MARL-based framework is trained through an actor-critic network and adopts the structure of centralized training and decentralized execution with partial observations. Comparing with model-based approaches, the developed MARL-based approach better captures the attributes of the distribution network, improves grid-level service performance, achieves better network constraints control, reduces the communication load, and achieves a faster response. The efficacy and efficiency of the developed method are verified by simulations on the IEEE 13-bus test feeder.
Read moreA Novel Hybrid Ledger Approach for Smart Grid Cyber Security and Energy Redistribution Using CryptoGrid
The increasing reliance on smart grid infrastructures, particularly for critical services such as hospitals, necessitates highly secure, efficient, and resilient energy management systems. In response to the vulnerabilities exposed during the COVID-19 pandemic, particularly in Gujarat where power failures disrupted ventilator operations, this paper introduces CryptoGrid, a novel hybrid cryptographic ledger framework for smart grid cyber security and energy redistribution. It is proposed to achieve grid security against advanced cyber threats using the proposed architecture that incorporates blockchain-powered identity authentication, federated learning, and GAN-based anomaly detection. The agent-based control systems and optimization of energy management with Multi-Agent Reinforcement Learning (MARL) framework with PSO-GA hybrid credit pricing and redistribution have been used to improve energy management. Priority nodes are critical infrastructures like hospitals and are provided in case of an outage automatically by mobile EV energy banks and the decentralized microgrid systems. The experimental findings indicate that CryptoGrid is efficient in providing secure and life-saving energy solutions with an anomaly detection accuracy of 98.7, a 72% drop in the energy wastage, and a 29% increase in emergency energy supply.
Read more3ML2MAS: a multi-agent reinforcement learning and BNNs-GAN integration framework for smart manufacturing optimization
ML2MAS: a multi-agent reinforcement learning and BNNs-GAN integration framework for smart manufacturing optimization
Adaptive Multi-Agent Reinforcement Learning for Dynamic Traffic Signal Optimization In Zero-Emission Urban Mobility
The high rates of urbanization and high concentration of vehicles have led to extreme traffic congestion, thereby contributing significantly to greenhouse gas emissions. This study proposes an Adaptive Multi-Agent Reinforcement Learning (MARL) framework combined with an Emission-Aware Reward System to optimize traffic lights. The suggested system in this work is decentralized MARL, where each intersection is considered as a single agent that disseminates the information to the neighboring agents to prevent congestion, coordinate with other agents, and attain an efficient traffic flow. This new dynamic signal timing algorithm, Traffic-Emission Adaptive Learning (TEAL), will examine actual traffic density, vehicle types, and other environmental statistics, such as the AAI (airquake index) and carbon footprint metric. The proposed module is an Eco-Mobility Prediction Module, a Graph Neural Network (GNN) predictor of congestion patterns and green wave synchronization, designed to reduce vehicle idle time and emissions. Furthermore, Vehicle-to-Infrastructure (V2I) is a mechanism that encourages the use of electric and hybrid vehicles, as it allows them to have a higher priority in zero-emission transportation. The proposed solution is shown to reduce CO2 emissions by 30 percent and is significantly more efficient than traditional models, as indicated by experimental findings conducted in a simulated urban setting. This new model, therefore, scales sustainable city and zero emission mobility up to an eco-friendly solution in the future smart city environment
Read more
Copyright 2026 Cactus Communications. All rights reserved.
