检索结果-内蒙古大学图书馆

Integrating Visible Light Communication and AI for Adaptive Traffic Management: A Focus on reward functions and Rerouting Coordination

引用

APPLIED SCIENCES-BASEL 2025年第1期15卷 116-116页

作者： Vieira, Manuela Galvao, Goncalo Vieira, Manuel A. Vestias, Mario Louro, Paula Vieira, Pedro DEETC ISEL IPL R Conselheiro Emidio Navarro P-1949014 Lisbon Portugal UNINOVA CTS P-2829516 Caparica Portugal LASI P-2829516 Caparica Portugal NOVA Sch Sci & Technol P-2829516 Caparica Portugal INOV INESC R Alves Redol 9 P-1000029 Lisbon Portugal Inst Super Tecn Inst Telecomunicacoes P-1049001 Lisbon Portugal

This study combines Visible Light Communication (VLC) and Artificial Intelligence (AI) to optimize traffic signal control, reduce congestion, and enhance safety. Utilizing existing road infrastructure, VLC technology transmits real-time data on vehicle and pedestrian positions, speeds, and queues. AI agents, powered by Deep Reinforcement Learning (DRL), process these data to manage traffic flows dynamically, applying anti-bottlenecking and rerouting techniques. A global agent coordinates local agents, enabling indirect communication and a unified DRL model that adjusts traffic light phases in real time using a queue/request/response system. A key focus of this work is the design of reward functions for standard and rerouting scenarios. In standard scenarios, the reward function prioritizes wide green bands for vehicles while penalizing pedestrian rule violations, balancing efficiency and safety. In rerouting scenarios, it dynamically prevents queuing spillovers at neighboring intersections, mitigating cascading congestion and ensuring safe, timely pedestrian crossings. Simulation experiments in the SUMO urban mobility simulator and real-world trials validate the system across diverse intersection types, including four-way crossings, T-intersections, and roundabouts. Results show significant reductions in vehicle and pedestrian waiting times, particularly in rerouting scenarios, demonstrating the system's scalability and adaptability. By integrating VLC technology and AI-driven adaptive control, this approach achieves efficient, safe, and flexible traffic management. The proposed system addresses urban mobility challenges effectively, offering a robust solution to modern traffic demands while improving the travel experience for all road users.

关键词： Visible Light Communication (VLC) Artificial Intelligence (AI) Deep Reinforcement Learning (DRL) traffic signal control traffic management reward functions rerouting scenarios pedestrian safety congestion reduction urban mobility adaptive traffic control

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of Different reward functions in Reinforcement Learning for the Scheduling of Modular Automotive Production Systems

引用

Procedia CIRP 2024年 126卷 81-86页

作者： Jan Markus Gelfgren Eric Tillmann Bill Tim Luther Simon Hagemann Sigrid Wenzel Mercedes-Benz Group AG Bela-Barenyi-Straße 1 71059 Sindelfingen Germany Department Organisation of Production and Factory Planning University of Kassel Kurt-Wolters-Straße 3 34125 Kassel Germany

Conventional, linear production lines struggle with the new flexibility requirements of the automotive market. Modular production has the potential to radically improve the production flexibility. However, scheduling modular production systems is still an open research question. Reinforcement learning (RL) is a form of artificial intelligence that shows great potential in scheduling complex modular production systems. Nonetheless, the performance of RL agents heavily depends on their reward function. Designing an optimal reward function is highly complex. This paper addresses this research gap and systematically compares six different reward functions for a variety of different modular production scenarios. In addition, a new learning method using resets is proposed and its performance is compared with the standard learning approach. The results suggest that dense reward functions perform better than a sparse one. However, there are major case-to-case discrepancies. The proposed learning method outperforms the standard learning method by 7 % on average. Nonetheless, the performance difference between different scenarios is larger than with the standard approach.

关键词： Reinforcement Learning reward functions Production Scheduling Modular Production

来源：评论

学校读者我要写书评

暂无评论

EPIC-Q: Equivalent-Policy Invariant Comparison Enhanced Transfer Q-learning for Run-Time SoC Performance-Power Optimization 24th

EPIC-Q: Equivalent-Policy Invariant Comparison Enhanced Tran...

引用

24th International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation

作者： Surhonne, Anmol Fawzi, Haitham S. Maurer, Florian Lenke, Oliver Meidinger, Michael Wild, Thomas Herkersdorf, Andreas Tech Univ Munich Arcistr 21 D-80333 Munich Germany

ISBN: (纸本)9783031783760;9783031783777

As power density becomes the main constraint of multicore systems, managing power consumption using DVFS while providing the desired performance becomes increasingly critical. Reinforcement learning (RL) performs significantly better than conventional methods in performance-power optimization under different hardware configurations and varying software applications. RL agents learn through trial-and-error by receiving rewards which is defined by an objective function (e.g. instructions-per-second (IPS)) within specified constraints (e.g. power budget). System and application requirements lead to changing objectives and constraints which in turn result in different reward functions. The RL agents adapt to these changing objectives and constraints (and hence reward functions). Equivalent-policy invariant comparison (EPIC) is a popular technique to evaluate different reward functions. EPIC provides a numerical score which quantifies the difference in two reward functions. In this work, we use this EPIC distance (score) to transfer knowledge and improve learning for changing reward functions. Experimental results using a DVFS enabled RISCV based system-on-chip implemented on an FPGA shows 16.2% lower power budget overshoots compared to a tabular Q-learning agent with direct transfer.

关键词： EPIC DVFS reinforcement learning reward functions power management transfer learning

来源：评论

学校读者我要写书评

暂无评论

Fine-tuning text-to-SQL models with reinforcement-learning training objectives

Natural Language Processing Journal

引用

Natural Language Processing Journal 2025年 10卷

作者： Xuan-Bang Nguyen Xuan-Hieu Phan Massimo Piccardi University of Engineering and Technology Vietnam National University Hanoi Viet Nam FPT Technology Research Institute FPT University Hanoi Viet Nam Faculty of Engineering and Information Technology University of Technology Sydney Broadway NSW 2007 Australia

Text-to-SQL is an important natural language processing task that helps users automatically convert natural language queries into formal SQL code. While transformer-based models have pushed text-to-SQL to unprecedented accuracy levels in recent years, such performance is confined to models of very large size that can only be run in specialised clouds. For this reason, in this paper we explore the use of reinforcement learning to improve the performance of models of more conservative size, which can fit within standard user hardware. As reinforcement learning reward, we propose a novel function which better aligns with the text-to-SQL evaluation metrics, applied in conjunction with two strong policy gradient algorithms, REINFORCE and RELAX. Our experimental results over the popular Spider benchmark show that the proposed approach has been able to outperform a conventionally-trained T5 Small baseline by 6.6 pp (percentage points) of exact-set-match accuracy and 4.6 pp of execution accuracy, and a T5 Base baseline by 2.0 pp and 1.9 pp, respectively. The proposed model has also achieved a remarkable comparative performance against ChatGPT instances.

关键词： Text-to-SQL Reinforcement learning reward functions Policy-gradient algorithms Fine-tuning

来源：评论

学校读者我要写书评

暂无评论

Adaptive traffic signal control system using composite reward architecture based deep reinforcement learning

引用

IET INTELLIGENT TRANSPORT SYSTEMS 2020年第14期14卷 2030-2041页

作者： Jamil, Abu Rafe Md Ganguly, Kishan Kumar Nower, Naushin Univ Dhaka Inst Informat Technol Dhaka 1000 Bangladesh

The increasing traffic congestion problem can be solved by an adaptive traffic signal control (ATSC) system as it utilises real-time traffic information to control traffic signals. Recently, deep reinforcement learning (DRL) has shown its potential in solving the traffic signal timing. However, one of the main challenges of DRL is to design a proper reward function and special attention needs for a multi-objective reward design. Since the feedback to the agent depends on the reward function, a proper design of reward function is needed for fast and stable learning. In this study, the authors proposed a new reward architecture called composite reward architecture (CRA) for multi-objective ATSC to optimise multiple objectives. It calculates multiple rewards in parallel for each action and applies the majority voting method to choose the desired action. Since the traffic signal of one intersection affects the adjacent intersections, a new coordination approach is proposed to get the overall smooth traffic flow. The proposed reward architecture CRA is compared with several existing reward functions used in the literature for different traffic scenarios. The new coordinated approach is compared with the non-coordinated approach. The authors demonstrated that the proposed approaches outperform the others concerning waiting time, halting the number of vehicles, and so on.

关键词： traffic engineering computing road traffic control optimisation adaptive control deep learning (artificial intelligence) deep reinforcement learning real-time traffic information DRL traffic signal timing multiobjective reward design fast learning stable learning composite reward architecture multiobjective ATSC multiple rewards coordination approach smooth traffic flow CRA reward functions traffic scenarios coordinated approach adaptive traffic signal control system traffic congestion problem majority voting method

来源：评论

学校读者我要写书评

暂无评论

Survey of Model-Based Reinforcement Learning: Applications on Robotics

引用

JOURNAL OF INTELLIGENT & ROBOTIC SYSTEMS 2017年第2期86卷 153-173页

作者： Polydoros, Athanasios S. Nalpantidis, Lazaros Aalborg Univ Dept Mech & Mfg Engn AC Meyers Vaenge 15 DK-2450 Copenhagen SV Denmark

Reinforcement learning is an appealing approach for allowing robots to learn new tasks. Relevant literature reveals a plethora of methods, but at the same time makes clear the lack of implementations for dealing with real life challenges. Current expectations raise the demand for adaptable robots. We argue that, by employing model-based reinforcement learning, the-now limited-adaptability characteristics of robotic systems can be expanded. Also, model-based reinforcement learning exhibits advantages that makes it more applicable to real life use-cases compared to model-free methods. Thus, in this survey, model-based methods that have been applied in robotics are covered. We categorize them based on the derivation of an optimal policy, the definition of the returns function, the type of the transition model and the learned task. Finally, we discuss the applicability of model-based reinforcement learning approaches in new applications, taking into consideration the state of the art in both algorithms and hardware.

关键词： Intelligent robotics Machine learning Model-based reinforcement learning Robot learning Policy search Transition models reward functions

来源：评论

学校读者我要写书评

暂无评论

Monte Carlo tree search control scheme for multibody dynamics applications

引用

NONLINEAR DYNAMICS 2024年第10期112卷 8363-8391页

作者： Tang, Yixuan Orzechowski, Grzegorz Prokop, Ales Mikkola, Aki LUT Univ Dept Mech Engn Lappeenranta 53850 Finland Brno Univ Technol Fac Mech Engn Technicka 2896-2 Brno 61669 Czech Republic

There is considerable interest in applying reinforcement learning (RL) to improve machine control across multiple industries, and the automotive industry is one of the prime examples. Monte Carlo Tree Search (MCTS) has emerged and proven powerful in decision-making games, even without understanding the rules. In this study, multibody system dynamics (MSD) control is first modeled as a Markov Decision Process and solved with Monte Carlo Tree Search. Based on randomized search space exploration, the MCTS framework builds a selective search tree by repeatedly applying a Monte Carlo rollout at each child node. However, without a library of available choices, deciding among the many possibilities for agent parameters can be intimidating. In addition, the MCTS poses a significant challenge for searching due to the large branching factor. This challenge is typically overcome by appropriate parameter design, search guiding, action reduction, parallelization, and early termination. To address these shortcomings, the overarching goal of this study is to provide needed insight into inverted pendulum controls via vanilla and modified MCTS agents, respectively. A series of reward functions are well-designed according to the control goal, which maps a specific distribution shape of reward bonus and guides the MCTS-based control to maintain the upright position. Numerical examples show that the reward-modified MCTS algorithms significantly improve the control performance and robustness of the default choice of a constant reward that constitutes the vanilla MCTS. The exponentially decaying reward functions perform better than the constant value or polynomial reward functions. Moreover, the exploitation vs. exploration trade-off and discount parameters are carefully tested. The study's results can guide the research of RL-based MSD users.

关键词： Monte Carlo Tree Search Multibody dynamics reward functions Parametric analysis Artificial intelligence control Inverted pendulum

来源：评论

学校读者我要写书评

暂无评论

Active reward learning with a novel acquisition function

引用

AUTONOMOUS ROBOTS 2015年第3期39卷 389-405页

作者： Daniel, Christian Kroemer, Oliver Viering, Malte Metz, Jan Peters, Jan Tech Univ Darmstadt D-64289 Darmstadt Germany Max Planck Inst Intelligente Syst D-72076 Tubingen Germany

reward functions are an essential component of many robot learning methods. Defining such functions, however, remains hard in many practical applications. For tasks such as grasping, there are no reliable success measures available. Defining reward functions by hand requires extensive task knowledge and often leads to undesired emergent behavior. We introduce a framework, wherein the robot simultaneously learns an action policy and a model of the reward function by actively querying a human expert for ratings. We represent the reward model using a Gaussian process and evaluate several classical acquisition functions (AFs) from the Bayesian optimization literature in this context. Furthermore, we present a novel AF, expected policy divergence. We demonstrate results of our method for a robot grasping task and show that the learned reward function generalizes to a similar task. Additionally, we evaluate the proposed novel AF on a real robot pendulum swing-up task.

关键词： Reinforcement learning Active learning Bayesian optimization Preference learning Inverse reinforcement learning reward functions Acquisition functions

来源：评论

学校读者我要写书评

暂无评论

Low complexity dynamic scheduling algorithm for real-time tasks

引用

ELECTRONICS LETTERS 1999年第24期35卷 2106-2108页

作者： Jung, G Kim, T Park, S Choi, K Ajou Univ Sch EE Suwon 442749 South Korea ETRI Taejon 305350 South Korea Ajou Univ Sch IE & CE Suwon 442749 South Korea

It is shown that the problem of maximising the total reward of online tasks can be solved by finding the minimum of the maximum derivatives of the reward functions. Based on the modified approach and a close observati... 详细信息

关键词： real-time tasks processor scheduling average complexity maximum derivatives task arrival characteristics online tasks dynamic scheduling algorithm reward functions Computational complexity heuristic algorithm real-time systems computational complexity

来源：评论

学校读者我要写书评

暂无评论

Performance evaluation with temporal rewards

引用

PERFORMANCE EVALUATION 2002年第2-3期50卷 189-218页

作者： Voeten, JPM Eindhoven Univ Technol Fac Elect Engn Eindhoven Embedded Syst Inst Sect Informat & Commun Syst NL-5600 MB Eindhoven Netherlands

Today many formalisms exist for specifying complex Markov chains. In contrast, formalisms for specifying rewards, enabling the analysis of long-run average performance properties, have remained quite primitive. Basically, they only support the analysis of relatively simple performance metrics that can be expressed as long-run averages of atomic rewards, i.e. rewards that are deductible directly from the individual states of the initial Markov chain specification. To deal with complex performance metrics that are dependent on the accumulation of atomic rewards over sequences of states, the initial specification has to be extended explicitly to provide the required state information. To solve this problem, we introduce in this paper a new formalism of temporal rewards that allows complex quantitative properties to be expressed in terms of temporal reward formulas. Together, an initial (discrete-time) Markov chain and the temporal reward formulas implicitly define an extended Markov chain that allows the determination of the quantitative property by traditional techniques for computing long-run averages. A method to construct the extended chain is given and it is proved that this method leaves long-run averages invariant for atomic rewards. We further establish conditions that guarantee the preservation of ergodicity. The construction method can build the extended chain in an on-the-fly manner allowing for efficient simulation. (C) 2002 Elsevier Science B.V. All rights reserved.

关键词： path-based reward variables reward functions temporal logic performance evaluation Markov chains

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：