UAV-Assisted Dynamic Avatar Task Migration for Vehicular Metaverse Services: A Multi-Agent Deep Reinforcement Learning Approach

Avatars, as promising digital representations and service assistants of users in Metaverses, can enable drivers and passengers to immerse themselves in 3D virtual services and spaces of UAV-assisted vehicular Metaverses. However, avatar tasks include a multitude of human-to-avatar and avatar-to-avatar interactive applications, e.g., augmented reality navigation,which consumes intensive computing resources. It is inefficient and impractical for vehicles to process avatar tasks locally. Fortunately, migrating avatar tasks to the nearest roadside units(RSU)or unmanned aerial vehicles(UAV) for execution is a promising solution to decrease computation overhead and reduce task processing latency, while the high mobility of vehicles brings challenges for vehicles to independently perform avatar migration decisions depending on current and future vehicle status. To address these challenges, in this paper, we propose a novel avatar task migration system based on multi-agent deep reinforcement learning(MADRL) to execute immersive vehicular avatar tasks dynamically. Specifically, we first formulate the problem of avatar task migration from vehicles to RSUs/UAVs as a partially observable Markov decision process that can be solved by MADRL algorithms. We then design the multi-agent proximal policy optimization(MAPPO) approach as the MADRL algorithm for the avatar task migration problem. To overcome slow convergence resulting from the curse of dimensionality and non-stationary issues caused by shared parameters in MAPPO, we further propose a transformer-based MAPPO approach via sequential decision-making models for the efficient representation of relationships among agents. Finally, to motivate terrestrial or non-terrestrial edge servers(e.g., RSUs or UAVs) to share computation resources and ensure traceability of the sharing records, we apply smart contracts and blockchain technologies to achieve secure sharing management. Numerical results demonstrate that the proposed approach outperforms the MAPPO approach by around 2% and effectively reduces approximately 20% of the latency of avatar task execution in UAV-assisted vehicular Metaverses.

Service Function Chain Deployment Algorithm Based on Multi-Agent Deep Reinforcement Learning

Aiming at the rapid growth of network services,which leads to the problems of long service request processing time and high deployment cost in the deployment of network function virtualization service function chain(SFC)under 5G networks,this paper proposes a multi-agent deep deterministic policy gradient optimization algorithm for SFC deployment(MADDPG-SD).Initially,an optimization model is devised to enhance the request acceptance rate,minimizing the latency and deploying the cost SFC is constructed for the network resource-constrained case.Subsequently,we model the dynamic problem as a Markov decision process(MDP),facilitating adaptation to the evolving states of network resources.Finally,by allocating SFCs to different agents and adopting a collaborative deployment strategy,each agent aims to maximize the request acceptance rate or minimize latency and costs.These agents learn strategies from historical data of virtual network functions in SFCs to guide server node selection,and achieve approximately optimal SFC deployment strategies through a cooperative framework of centralized training and distributed execution.Experimental simulation results indicate that the proposed method,while simultaneously meeting performance requirements and resource capacity constraints,has effectively increased the acceptance rate of requests compared to the comparative algorithms,reducing the end-to-end latency by 4.942%and the deployment cost by 8.045%.

Multi-Agent Deep Reinforcement Learning for Efficient Computation Offloading in Mobile Edge Computing

Mobile-edge computing(MEC)is a promising technology for the fifth-generation(5G)and sixth-generation(6G)architectures,which provides resourceful computing capabilities for Internet of Things(IoT)devices,such as virtual reality,mobile devices,and smart cities.In general,these IoT applications always bring higher energy consumption than traditional applications,which are usually energy-constrained.To provide persistent energy,many references have studied the offloading problem to save energy consumption.However,the dynamic environment dramatically increases the optimization difficulty of the offloading decision.In this paper,we aim to minimize the energy consumption of the entireMECsystemunder the latency constraint by fully considering the dynamic environment.UnderMarkov games,we propose amulti-agent deep reinforcement learning approach based on the bi-level actorcritic learning structure to jointly optimize the offloading decision and resource allocation,which can solve the combinatorial optimization problem using an asymmetric method and compute the Stackelberg equilibrium as a better convergence point than Nash equilibrium in terms of Pareto superiority.Our method can better adapt to a dynamic environment during the data transmission than the single-agent strategy and can effectively tackle the coordination problem in the multi-agent environment.The simulation results show that the proposed method could decrease the total computational overhead by 17.8%compared to the actor-critic-based method and reduce the total computational overhead by 31.3%,36.5%,and 44.7%compared with randomoffloading,all local execution,and all offloading execution,respectively.

UAV Frequency-based Crowdsensing Using Grouping Multi-agent Deep Reinforcement Learning

Mobile CrowdSensing(MCS)is a promising sensing paradigm that recruits users to cooperatively perform sensing tasks.Recently,unmanned aerial vehicles(UAVs)as the powerful sensing devices are used to replace user participation and carry out some special tasks,such as epidemic monitoring and earthquakes rescue.In this paper,we focus on scheduling UAVs to sense the task Point-of-Interests(PoIs)with different frequency coverage requirements.To accomplish the sensing task,the scheduling strategy needs to consider the coverage requirement,geographic fairness and energy charging simultaneously.We consider the complex interaction among UAVs and propose a grouping multi-agent deep reinforcement learning approach(G-MADDPG)to schedule UAVs distributively.G-MADDPG groups all UAVs into some teams by a distance-based clustering algorithm(DCA),then it regards each team as an agent.In this way,G-MADDPG solves the problem that the training time of traditional MADDPG is too long to converge when the number of UAVs is large,and the trade-off between training time and result accuracy could be controlled flexibly by adjusting the number of teams.Extensive simulation results show that our scheduling strategy has better performance compared with three baselines and is flexible in balancing training time and result accuracy.

Multi-Agent Deep Reinforcement Learning for Cross-Layer Scheduling in Mobile Ad-Hoc Networks

Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus on enabling congestion control to minimize network transmission delays through flexible power control.To effectively solve the congestion problem,we propose a distributed cross-layer scheduling algorithm,which is empowered by graph-based multi-agent deep reinforcement learning.The transmit power is adaptively adjusted in real-time by our algorithm based only on local information(i.e.,channel state information and queue length)and local communication(i.e.,information exchanged with neighbors).Moreover,the training complexity of the algorithm is low due to the regional cooperation based on the graph attention network.In the evaluation,we show that our algorithm can reduce the transmission delay of data flow under severe signal interference and drastically changing channel states,and demonstrate the adaptability and stability in different topologies.The method is general and can be extended to various types of topologies.

Exploring Local Chemical Space in De Novo Molecular Generation Using Multi-Agent Deep Reinforcement Learning

Single-agent reinforcement learning (RL) is commonly used to learn how to play computer games, in which the agent makes one move before making the next in a sequential decision process. Recently single agent was also employed in the design of molecules and drugs. While a single agent is a good fit for computer games, it has limitations when used in molecule design. Its sequential learning makes it impossible to modify or improve the previous steps while working on the current step. In this paper, we proposed to apply the multi-agent RL approach to the research of molecules, which can optimize all sites of a molecule simultaneously. To elucidate the validity of our approach, we chose one chemical compound Favipiravir to explore its local chemical space. Favipiravir is a broad-spectrum inhibitor of viral RNA polymerase, and is one of the compounds that are currently being used in SARS-CoV-2 (COVID-19) clinical trials. Our experiments revealed the collaborative learning of a team of deep RL agents as well as the learning of its individual learning agent in the exploration of Favipiravir. In particular, our multi-agents not only discovered the molecules near Favipiravir in chemical space, but also the learnability of each site in the string representation of Favipiravir, critical information for us to understand the underline mechanism that supports machine learning of molecules.

Multi-Agent Deep Reinforcement Learning-Based Resource Allocation in HPC/AI Converged Cluster

As the complexity of deep learning(DL)networks and training data grows enormously,methods that scale with computation are becoming the future of artificial intelligence(AI)development.In this regard,the interplay between machine learning(ML)and high-performance computing(HPC)is an innovative paradigm to speed up the efficiency of AI research and development.However,building and operating an HPC/AI converged system require broad knowledge to leverage the latest computing,networking,and storage technologies.Moreover,an HPC-based AI computing environment needs an appropriate resource allocation and monitoring strategy to efficiently utilize the system resources.In this regard,we introduce a technique for building and operating a high-performance AI-computing environment with the latest technologies.Specifically,an HPC/AI converged system is configured inside Gwangju Institute of Science and Technology(GIST),called GIST AI-X computing cluster,which is built by leveraging the latest Nvidia DGX servers,high-performance storage and networking devices,and various open source tools.Therefore,it can be a good reference for building a small or middlesized HPC/AI converged system for research and educational institutes.In addition,we propose a resource allocation method for DL jobs to efficiently utilize the computing resources with multi-agent deep reinforcement learning(mDRL).Through extensive simulations and experiments,we validate that the proposed mDRL algorithm can help the HPC/AI converged cluster to achieve both system utilization and power consumption improvement.By deploying the proposed resource allocation method to the system,total job completion time is reduced by around 20%and inefficient power consumption is reduced by around 40%.

Real-time Operation Optimization in Active Distribution Networks Based on Multi-agent Deep Reinforcement Learning

The increasing integration of intermittent renewable energy sources(RESs)poses great challenges to active distribution networks(ADNs),such as frequent voltage fluctuations.This paper proposes a novel ADN strategy based on multiagent deep reinforcement learning(MADRL),which harnesses the regulating function of switch state transitions for the realtime voltage regulation and loss minimization.After deploying the calculated optimal switch topologies,the distribution network operator will dynamically adjust the distributed energy resources(DERs)to enhance the operation performance of ADNs based on the policies trained by the MADRL algorithm.Owing to the model-free characteristics and the generalization of deep reinforcement learning,the proposed strategy can still achieve optimization objectives even when applied to similar but unseen environments.Additionally,integrating parameter sharing(PS)and prioritized experience replay(PER)mechanisms substantially improves the strategic performance and scalability.This framework has been tested on modified IEEE 33-bus,IEEE 118-bus,and three-phase unbalanced 123-bus systems.The results demonstrate the significant real-time regulation capabilities of the proposed strategy.

Discovering Latent Variables for the Tasks With Confounders in Multi-Agent Reinforcement Learning

Efficient exploration in complex coordination tasks has been considered a challenging problem in multi-agent reinforcement learning(MARL). It is significantly more difficult for those tasks with latent variables that agents cannot directly observe. However, most of the existing latent variable discovery methods lack a clear representation of latent variables and an effective evaluation of the influence of latent variables on the agent. In this paper, we propose a new MARL algorithm based on the soft actor-critic method for complex continuous control tasks with confounders. It is called the multi-agent soft actor-critic with latent variable(MASAC-LV) algorithm, which uses variational inference theory to infer the compact latent variables representation space from a large amount of offline experience.Besides, we derive the counterfactual policy whose input has no latent variables and quantify the difference between the actual policy and the counterfactual policy via a distance function. This quantified difference is considered an intrinsic motivation that gives additional rewards based on how much the latent variable affects each agent. The proposed algorithm is evaluated on two collaboration tasks with confounders, and the experimental results demonstrate the effectiveness of MASAC-LV compared to other baseline algorithms.

Unleashing the Power of Multi-Agent Reinforcement Learning for Algorithmic Trading in the Digital Financial Frontier and Enterprise Information Systems

In the rapidly evolving landscape of today’s digital economy,Financial Technology(Fintech)emerges as a trans-formative force,propelled by the dynamic synergy between Artificial Intelligence(AI)and Algorithmic Trading.Our in-depth investigation delves into the intricacies of merging Multi-Agent Reinforcement Learning(MARL)and Explainable AI(XAI)within Fintech,aiming to refine Algorithmic Trading strategies.Through meticulous examination,we uncover the nuanced interactions of AI-driven agents as they collaborate and compete within the financial realm,employing sophisticated deep learning techniques to enhance the clarity and adaptability of trading decisions.These AI-infused Fintech platforms harness collective intelligence to unearth trends,mitigate risks,and provide tailored financial guidance,fostering benefits for individuals and enterprises navigating the digital landscape.Our research holds the potential to revolutionize finance,opening doors to fresh avenues for investment and asset management in the digital age.Additionally,our statistical evaluation yields encouraging results,with metrics such as Accuracy=0.85,Precision=0.88,and F1 Score=0.86,reaffirming the efficacy of our approach within Fintech and emphasizing its reliability and innovative prowess.

Regional Multi-Agent Cooperative Reinforcement Learning for City-Level Traffic Grid Signal Control

This article studies the effective traffic signal control problem of multiple intersections in a city-level traffic system.A novel regional multi-agent cooperative reinforcement learning algorithm called RegionSTLight is proposed to improve the traffic efficiency.Firstly a regional multi-agent Q-learning framework is proposed,which can equivalently decompose the global Q value of the traffic system into the local values of several regions Based on the framework and the idea of human-machine cooperation,a dynamic zoning method is designed to divide the traffic network into several strong-coupled regions according to realtime traffic flow densities.In order to achieve better cooperation inside each region,a lightweight spatio-temporal fusion feature extraction network is designed.The experiments in synthetic real-world and city-level scenarios show that the proposed RegionS TLight converges more quickly,is more stable,and obtains better asymptotic performance compared to state-of-theart models.

Collision-free parking recommendation based on multi-agent reinforcement learning in vehicular crowdsensing

The recent proliferation of Fifth-Generation(5G)networks and Sixth-Generation(6G)networks has given rise to Vehicular Crowd Sensing(VCS)systems which solve parking collisions by effectively incentivizing vehicle participation.However,instead of being an isolated module,the incentive mechanism usually interacts with other modules.Based on this,we capture this synergy and propose a Collision-free Parking Recommendation(CPR),a novel VCS system framework that integrates an incentive mechanism,a non-cooperative VCS game,and a multi-agent reinforcement learning algorithm,to derive an optimal parking strategy in real time.Specifically,we utilize an LSTM method to predict parking areas roughly for recommendations accurately.Its incentive mechanism is designed to motivate vehicle participation by considering dynamically priced parking tasks and social network effects.In order to cope with stochastic parking collisions,its non-cooperative VCS game further analyzes the uncertain interactions between vehicles in parking decision-making.Then its multi-agent reinforcement learning algorithm models the VCS campaign as a multi-agent Markov decision process that not only derives the optimal collision-free parking strategy for each vehicle independently,but also proves that the optimal parking strategy for each vehicle is Pareto-optimal.Finally,numerical results demonstrate that CPR can accomplish parking tasks at a 99.7%accuracy compared with other baselines,efficiently recommending parking spaces.

Safety-Constrained Multi-Agent Reinforcement Learning for Power Quality Control in Distributed Renewable Energy Networks

This paper examines the difficulties of managing distributed power systems,notably due to the increasing use of renewable energy sources,and focuses on voltage control challenges exacerbated by their variable nature in modern power grids.To tackle the unique challenges of voltage control in distributed renewable energy networks,researchers are increasingly turning towards multi-agent reinforcement learning(MARL).However,MARL raises safety concerns due to the unpredictability in agent actions during their exploration phase.This unpredictability can lead to unsafe control measures.To mitigate these safety concerns in MARL-based voltage control,our study introduces a novel approach:Safety-ConstrainedMulti-Agent Reinforcement Learning(SC-MARL).This approach incorporates a specialized safety constraint module specifically designed for voltage control within the MARL framework.This module ensures that the MARL agents carry out voltage control actions safely.The experiments demonstrate that,in the 33-buses,141-buses,and 322-buses power systems,employing SC-MARL for voltage control resulted in a reduction of the Voltage Out of Control Rate(%V.out)from0.43,0.24,and 2.95 to 0,0.01,and 0.03,respectively.Additionally,the Reactive Power Loss(Q loss)decreased from 0.095,0.547,and 0.017 to 0.062,0.452,and 0.016 in the corresponding systems.

A survey on multi-agent reinforcement learning and its application

Multi-agent reinforcement learning(MARL)has been a rapidly evolving field.This paper presents a comprehensive survey of MARL and its applications.We trace the historical evolution of MARL,highlight its progress,and discuss related survey works.Then,we review the existing works addressing inherent challenges and those focusing on diverse applications.Some representative stochastic games,MARL means,spatial forms of MARL,and task classification are revisited.We then conduct an in-depth exploration of a variety of challenges encountered in MARL applications.We also address critical operational aspects,such as hyperparameter tuning and computational complexity,which are pivotal in practical implementations of MARL.Afterward,we make a thorough overview of the applications of MARL to intelligent machines and devices,chemical engineering,biotechnology,healthcare,and societal issues,which highlights the extensive potential and relevance of MARL within both current and future technological contexts.Our survey also encompasses a detailed examination of benchmark environments used in MARL research,which are instrumental in evaluating MARL algorithms and demonstrate the adaptability of MARL to diverse application scenarios.In the end,we give our prospect for MARL and discuss their related techniques and potential future applications.

Performance Evaluation ofMulti-Agent Reinforcement Learning Algorithms

Multi-Agent Reinforcement Learning(MARL)has proven to be successful in cooperative assignments.MARL is used to investigate how autonomous agents with the same interests can connect and act in one team.MARL cooperation scenarios are explored in recreational cooperative augmented reality environments,as well as realworld scenarios in robotics.In this paper,we explore the realm of MARL and its potential applications in cooperative assignments.Our focus is on developing a multi-agent system that can collaborate to attack or defend against enemies and achieve victory withminimal damage.To accomplish this,we utilize the StarCraftMulti-Agent Challenge(SMAC)environment and train four MARL algorithms:Q-learning with Mixtures of Experts(QMIX),Value-DecompositionNetwork(VDN),Multi-agent Proximal PolicyOptimizer(MAPPO),andMulti-Agent Actor Attention Critic(MAA2C).These algorithms allow multiple agents to cooperate in a specific scenario to achieve the targeted mission.Our results show that the QMIX algorithm outperforms the other three algorithms in the attacking scenario,while the VDN algorithm achieves the best results in the defending scenario.Specifically,the VDNalgorithmreaches the highest value of battle wonmean and the lowest value of dead alliesmean.Our research demonstrates the potential forMARL algorithms to be used in real-world applications,such as controllingmultiple robots to provide helpful services or coordinating teams of agents to accomplish tasks that would be impossible for a human to do.The SMAC environment provides a unique opportunity to test and evaluate MARL algorithms in a challenging and dynamic environment,and our results show that these algorithms can be used to achieve victory with minimal damage.

Automatic depth matching method of well log based on deep reinforcement learning

In the traditional well log depth matching tasks,manual adjustments are required,which means significantly labor-intensive for multiple wells,leading to low work efficiency.This paper introduces a multi-agent deep reinforcement learning(MARL)method to automate the depth matching of multi-well logs.This method defines multiple top-down dual sliding windows based on the convolutional neural network(CNN)to extract and capture similar feature sequences on well logs,and it establishes an interaction mechanism between agents and the environment to control the depth matching process.Specifically,the agent selects an action to translate or scale the feature sequence based on the double deep Q-network(DDQN).Through the feedback of the reward signal,it evaluates the effectiveness of each action,aiming to obtain the optimal strategy and improve the accuracy of the matching task.Our experiments show that MARL can automatically perform depth matches for well-logs in multiple wells,and reduce manual intervention.In the application to the oil field,a comparative analysis of dynamic time warping(DTW),deep Q-learning network(DQN),and DDQN methods revealed that the DDQN algorithm,with its dual-network evaluation mechanism,significantly improves performance by identifying and aligning more details in the well log feature sequences,thus achieving higher depth matching accuracy.

Optimal Secondary Control of Islanded AC Microgrids with Communication Time-delay Based on Multi-agent Deep Reinforcement Learning

In this paper,an optimal secondary control strategy is proposed for islanded AC microgrids considering communi-cation time-delays.The proposed method is designed based on the data-driven principle,which consists of an offine training phase and online application phase.For offline training,each control agent is formulated by a deep neural network(DNN)and trained based on a multi-agent deep reinforcement learning(MA-DRL)framework.A deep deterministic policy gradient(DDPG)algorithm is improved and applied to search for an optimal policy of the secondary control,where a global cost function is developed to evaluate the overall control performance.In addition,the communication time-delay is introduced in the system to enrich training scenarios,which aims to solve the time-delay problem in the secondary control.For the online stage,each controller is deployed in a distributed way which only requires local and neighboring information for each DG.Based on this,the well-trained controllers can provide optimal solutions under load variations,and communication time-delays for online applications.Several case studies are conducted to validate the feasibility and stability of the proposed secondary control.Index Terms-Communication time-delay,global cost function,islanded AC microgrid,multi-agent deep reinforcement learning(MA-DRL),secondary control.

A multi-agent deep reinforcement learning approach for solving the multi-depot vehicle routing problem

The multi-depot vehicle routing problem(MDVRP)is one of the most essential and useful variants of the traditional vehicle routing problem(VRP)in supply chain management(SCM)and logistics studies.Many supply chains(SC)choose the joint distribution of multiple depots to cut transportation costs and delivery times.However,the ability to deliver quality and fast solutions for MDVRP remains a challenging task.Traditional optimization approaches in operation research(OR)may not be practical to solve MDVRP in real-time.With the latest developments in artificial intelligence(AI),it becomes feasible to apply deep reinforcement learning(DRL)for solving combinatorial routing problems.This paper proposes a new multi-agent deep reinforcement learning(MADRL)model to solve MDVRP.Extensive experiments are conducted to evaluate the performance of the proposed approach.Results show that the developed MADRL model can rapidly capture relative information embedded in graphs and effectively produce quality solutions in real-time.

Multi-energy Management of Interconnected Multi-microgrid System Using Multi-agent Deep Reinforcement Learning

The multi-directional flow of energy in a multi-microgrid(MMG) system and different dispatching needs of multiple energy sources in time and location hinder the optimal operation coordination between microgrids. We propose an approach to centrally train all the agents to achieve coordinated control through an individual attention mechanism with a deep dense neural network for reinforcement learning. The attention mechanism and novel deep dense neural network allow each agent to attend to the specific information that is most relevant to its reward. When training is complete, the proposed approach can construct decisions to manage multiple energy sources within the MMG system in a fully decentralized manner. Using only local information, the proposed approach can coordinate multiple internal energy allocations within individual microgrids and external multilateral multi-energy interactions among interconnected microgrids to enhance the operational economy and voltage stability. Comparative results demonstrate that the cost achieved by the proposed approach is at most 71.1% lower than that obtained by other multi-agent deep reinforcement learning approaches.

A deep reinforcement learning approach to gasoline blending real-time optimization under uncertainty

The gasoline inline blending process has widely used real-time optimization techniques to achieve optimization objectives,such as minimizing the cost of production.However,the effectiveness of real-time optimization in gasoline blending relies on accurate blending models and is challenged by stochastic disturbances.Thus,we propose a real-time optimization algorithm based on the soft actor-critic(SAC)deep reinforcement learning strategy to optimize gasoline blending without relying on a single blending model and to be robust against disturbances.Our approach constructs the environment using nonlinear blending models and feedstocks with disturbances.The algorithm incorporates the Lagrange multiplier and path constraints in reward design to manage sparse product constraints.Carefully abstracted states facilitate algorithm convergence,and the normalized action vector in each optimization period allows the agent to generalize to some extent across different target production scenarios.Through these well-designed components,the algorithm based on the SAC outperforms real-time optimization methods based on either nonlinear or linear programming.It even demonstrates comparable performance with the time-horizon based real-time optimization method,which requires knowledge of uncertainty models,confirming its capability to handle uncertainty without accurate models.Our simulation illustrates a promising approach to free real-time optimization of the gasoline blending process from uncertainty models that are difficult to acquire in practice.