检索结果-内蒙古大学图书馆

Reinforcement learning-based control with application to the once-through steam generator system

NUCLEAR ENGINEERING AND TECHNOLOGY 2023年第10期55卷 3515-3524页

作者： Li, Cheng Yu, Ren Yu, Wenmin Wang, Tianshu Naval Univ Engn Wuhan 430033 Peoples R China

A reinforcement learning framework is proposed for the control problem of outlet steam pressure of the once-through steam generator(OTSG) in this paper. The double-layer controller using Proximal Policy Optimization(ppo) algorithm is applied in the control structure of the OTSG. The ppo algorithm can train the neural networks continuously according to the process of interaction with the environment and then the trained controller can realize better control for the OTSG. Meanwhile, reinforcement learning has the characteristic of difficult application in real-world objects, this paper proposes an innovative pretraining method to solve this problem. The difficulty in the application of reinforcement learning lies in training. The optimal strategy of each step is summed up through trial and error, and the training cost is very high. In this paper, the LSTM model is adopted as the training environment for pretraining, which saves training time and improves efficiency. The experimental results show that this method can realize the self-adjustment of control parameters under various working conditions, and the control effect has the advantages of small overshoot, fast stabilization speed, and strong adaptive ability. (c) 2023 Korean Nuclear Society, Published by Elsevier Korea LLC. This is an open access article under the CC BY-NC-ND license (http://***/licenses/by-nc-nd/4.0/).

关键词： LSTM ppo algorithm Pretraining Once-through steam generator

来源：评论

学校读者我要写书评

暂无评论

Enhancing Mario Gaming Using Optimized Reinforcement Learning 20th

Enhancing Mario Gaming Using Optimized Reinforcement Learnin...

引用

20th International Conference on Distributed Computing and Intelligent Technology (ICDCIT)

作者： Sah, Sumit Kumar Fidele, Hategekimana NITTE Meenakshu Inst Technol Bangalore 560064 India Univ Cent Africa Dept Adventist Gishushu Campus Kigali Rwanda

ISBN: (纸本)9783031505829;9783031505836

"In the realm of classic gaming, Mario has held a special place in the hearts of players for generations. This study, titled 'Enhancing Mario Gaming using Optimized Reinforcement Learning', ventures into the uncharted territory of machine learning to elevate the Mario gaming experience to new heights. Our research employs state-of-the-art techniques, including the Proximal Policy Optimization (ppo) algorithm and Convolutional Neural Networks (CNN), to infuse intelligence into the Mario gameplay. By optimizing reinforcement learning, we aim to create an immersive and engaging experience for players. In addition to the technical aspects, we delve into the concept of game appeal, a pivotal component in capturing player engagement. Our innovative approach blends the prowess of ppo, CNN, and reinforcement learning to unlock unique insights and methodologies for enhancing Mario games. This comprehensive analysis provides actionable guidance for selecting the most suitable techniques for distinct facets of Mario games. The culmination is an enriched, captivating, and optimized gaming experience that befits the title, 'Enhancing Mario Gaming using Optimized Reinforcement Learning'.

关键词： Mario games Reinforcement learning ppo algorithm CNN Game enhancement Player engagement

来源：评论

学校读者我要写书评

暂无评论

Energy Storage Scheduling Optimization Strategy Based on Deep Reinforcement Learning 18th

Energy Storage Scheduling Optimization Strategy Based on Dee...

引用

18th International Conference on Neural Networks (ISNN)

作者： Hou, Shixi Han, Jienan Liu, Xiangjiang Guo, Ruoshan Chu, Yundi Hohai Univ Coll Artificial Intelligence & Automat Changzhou 213000 Peoples R China

ISBN: (纸本)9789819743988;9789819743995

Renewable energy growth will be a top priority for China's future energy development. However, while vigorously developing renewable energy, the problem of curtailment of wind and solar power has also arisen. In order to make full use of renewable energy, this paper constructs an energy storage scheduling model based on deep intensive chemical Xi. Since most of the index parameters in the actual complex scenario are continuous variables, in order to better simulate the real situation, this paper proposes a hybrid energy storage model based on the blocked energy in the continuous action space, and further incorporates solar energy on the basis of considering wind energy. At the same time, the model also introduces an alkaline water electrolyzer to convert electrical energy into hydrogen for storage, which enriches the destination and storage form of electric energy, and also makes the profit mode of electric field more diversified. It is worth mentioning that the model also takes into account the maintenance costs and related financial costs of the equipment when calculating the benefits and costs, so that the model is closer to the real production and life. By comparing the similarities and differences between the two in the training process and test results, the feasibility of energy storage scheduling in the face of complex scenarios is verified.

关键词： Hybrid energy storage model Alkaline water electrolyzers TD3 algorithm ppo algorithm

来源：评论

学校读者我要写书评

暂无评论

Gain Scheduled PI controller design using Multi-Objective Reinforcement Learning 4

Gain Scheduled PI controller design using Multi-Objective Re...

引用

4th IFAC Conference on Advances in Proportional-Integral-Derivate Control (PID)

作者： Kranthi, Kumar P. Detroja, Ketan P. Indian Inst Technol Dept Artificial Intelligence Hyderabad Telangana India

Gain scheduling, a widely adopted control design approach, performs tasks by decomposing into sub-problems. It has found successful applications across diverse fields, including aerospace and industrial process control (Leith and Leithead (2000)). Reinforcement Learning (RL) has become increasingly significant in process control applications, thanks to the advancement of sophisticated algorithms and the rise in computational speeds. An attempt has been made to enhance the performance of a Proportional-Integral (PI) controller by integrating the benefits of both gain scheduling and RL. The manuscript focuses on designing a Gain Scheduled PI controller using Multi-Objective RL (GS-MORL). Proposing the training of multiple RL-based PI controllers with distinct objectives with customized reward functions to derive tuning parameters for each goal. The switching between these parameter sets is facilitated by a gain schedule variable. The controller parameters are adjusted based on the error between actual and desired responses. The aim is to enhance overall performance across various performance indices, including the integral of absolute error and time response specifications. The performance improvement from the proposed work has been demonstrated using two simulation case studies, one involving an underdamped second-order system and an integrating type steam turbine system. Comparative analysis against existing literature revealed superior performance across all indices. Copyright (C) 2024 The Authors. This is an open access article under the CC BY-NC-ND license (https://***/licenses/by-nc-nd/4.0/)

关键词： Gain Scheduling Multi-Objective RL PI controller ppo algorithm Steam Turbine System Step response

来源：评论

学校读者我要写书评

暂无评论

Eco-Driving and Suppression of Traffic Waves in Mixed Heterogeneous Traffic Streams Using Connected-Automated Vehicle

Eco-Driving and Suppression of Traffic Waves in Mixed Hetero...

引用

International Conference on Artificial Intelligence and Autonomous Transportation, AIAT 2024

作者： Gu, Yunyang Xu, Yueru College of Automation Jiangsu University of Science and Technology Zhenjiang212100 China Intelligent Transportation System Research Center School of Transportation Jiangsu Key Laboratory of Urban ITS Southeast University Nanjing211189 China

ISBN: (纸本)9789819639564

Due to human driving habits, traffic flow often has stop-and-go phenomena, resulting in reduced traffic efficiency, increased fuel consumption and emissions. This paper models the general heterogeneous mixed traffic flow and describes in detail the driving characteristics of human drivers in the traffic flow. Based on this model, an ecological driving strategy for autonomous vehicles is proposed. The optimization problem of autonomous vehicles is established based on the heterogeneous mixed traffic flow model. An improved proximal policy optimization (ppo) algorithm is designed to solve the optimal control input of the grid-connected autonomous vehicle (CAV). The effectiveness of the optimization algorithm is demonstrated through numerical simulation. © Beijing Paike Culture Commu. Co., Ltd. 2025.

关键词： Autonomous Driving Eco-efficient Driving Mixed Traffic Flow ppo algorithm Sliding Mode Control

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Empowered Deep Network Enabled Optimal Controlled Environment For Crop Yield 15

Reinforcement Empowered Deep Network Enabled Optimal Control...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Gorantla, Snehitha Veluchamy, S. Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering - Artificial Intelligence Chennai601103 India Amrita Vishwa Vidyapeetham Amrita School of Engineering Department of Electronics and Communication Engineering Chennai601103 India

ISBN: (纸本)9798350370249

This paper focuses on developing a comprehensive system for optimised climate control in a simulated greenhouse setting with complex environment modelling, communication-enabled agents, and hyperparameter optimisation using multiagent reinforcement learning (MARL). The greenhouse environment model gives the agents a richer and more detailed state representation by extending pertinent climate variables, plant health indicators, and resource levels for each zone. In order to stimulate both individual zone performance and overall greenhouse goals, the incentive function has been changed. This involves goals like preserving a stable climate, optimising resource use, or achieving a balanced yield. Based on the communication-based MARL strategy used, information sharing between agents is made possible by the communication protocols included into multi-agent architecture. This enables the agents to work together and make decisions in unison. Neural network designs are used to build both value and policy networks, allowing the agents to understand intricate relationships and decide on the best course of action depending on the observed situation. © 2024 IEEE.

关键词： Crop Yield Greenhouse Multi-Agent Systems Neural Network ppo algorithm Reinforcement Learning Resource Management

来源：评论

学校读者我要写书评

暂无评论

Design of Reinforcement Learning based PI controller for nonlinear Multivariable System

Design of Reinforcement Learning based PI controller for non...

引用

European Control Conference (ECC)

作者： Kumar, Kranthi P. Detroja, Ketan P. Indian Inst Technol Dept Artificial Intelligence Hyderabad India

ISBN: (纸本)9783907144084

With the advancement of computational power, it has become easier to approximate a complex policy function using a deep neural network to achieve better accuracy and performance. Hence the policies are formulated as parameterized deep neural networks and the same are trained using state-of-the-art policy gradient algorithms. Therefore in recent years, reinforcement learning has gained much attention and its advantages make it ideal for the adaptive tuning of PI controllers. This manuscript aims to build an adaptive PI controller using Reinforcement Learning (RL) and Deep Neural Networks (DNN) for 1) nonlinear multivariable systems and 2) non-minimum phase multivariable systems. The main objective is to design a PI controller using RL formulation for nonlinear processes, where nonlinear dynamics may be unknown, by training the agents through interactions with linearized models. Proximal Policy Optimization (ppo) algorithm with dynamic action space is proposed in this manuscript. The proposed ppo based controllers achieve better performance and provide smoother convergence compared to the existing RL based controller. The effectiveness of the proposed RL-based PI controller is demonstrated through a Quadruple-Tank system with nonlinear dynamic behavior. The system exhibits both minimum and non-minimum characteristics based on the multivariable zero location. The reward function is also adapted to aid the control applications and leads to efficient learning and convergence. The results show that the controller performed well on the actual process in both minimum and non-minimum phase cases.

关键词： PI controller ppo algorithm Quadruple-Tank multivariable system Minimum and Non-minimum Phase systems

来源：评论

学校读者我要写书评

暂无评论

Automated Crop Growth Monitoring and Optimizing the Yield with Reinforcement Learning 3

Automated Crop Growth Monitoring and Optimizing the Yield wi...

引用

3rd IEEE International Conference on Mobile Networks and Wireless Communications, ICMNWC 2023

作者： Gorantla, Snehitha Veluchamy, S. Amrita School of Computing Department of Computer Science and Engineering - Artificial Intelligence Amrita Vishwa Vidyapeetham Chennai601103 India Amrita School of Computing Department of Computer Science and Engineering - Cyber Security Amrita Vishwa Vidyapeetham Chennai601103 India

ISBN: (纸本)9798350317022

A reinforcement learning agent for optimal green-house management through Proximal Policy Optimisation is developed and evaluated in this work. With variable elements like lighting, irrigation, humidity, and temperature, a virtual greenhouse is produced. By interacting with the surroundings, the agent modifies the controls and gets input on things like crop health and growth rate. For policy and value estimation, the neural network design consists of distinct actor and critic networks as well as shared feature layers. Compared to fixed or random policies, the trained agent shows the ability to regulate conditions for increased crop yield and sustainability. As the number of trials increased, the RL agent's average return per trial increased steadily. This suggests that the RL agent has become more adept at controlling the greenhouse environment in subsequent experiments. This demonstrates how well the RL agent can manage crop health and resource use in the greenhouse climate control system. © 2023 IEEE.

关键词： Crop Yield Greenhouse Neural Network ppo algorithm Reinforcement Learning

来源：评论

学校读者我要写书评

暂无评论

Fixed-Wing Stalled Maneuver Control Technology Based on Deep Reinforcement Learning 5

Fixed-Wing Stalled Maneuver Control Technology Based on Deep...

引用

5th IEEE International Conference on Big Data and Artificial Intelligence (BDAI)

作者： Hu, Weijun Gao, Zhiqiang Quan, Jiale Ma, Xianlong Xiong, Jingyi Zhang, Weijie Northwestern Polytech Univ Sch Astronaut Xian Peoples R China

ISBN: (数字)9781665470810

ISBN: (纸本)9781665470810

A fixed-wing flight control method based on Deep Reinforcement Learning (DRL) was proposed to solve the problem of strong coupling between the channels of a fixed-wing aircraft during actual flight control, and the strong non-linearity and uncertainty of the aerodynamic parameters of the aircraft during stalled maneuver. When designing the control system, Proximal Policy Optimization (ppo) and neural network are used to design a control method that is directly mapped from the state to the aircraft actuator. A formalized reward function is designed to simulate the control of the vehicle's stalled maneuver, taking into account the relationship between the position and attitude of the vehicle and the actual demand command. The simulation results show that the control system design based on ppo algorithm can reduce the dependence on the model, achieve the intelligent control of the aircraft.

关键词： DRL ppo algorithm fixed-wing formalized reward stalled maneuver

来源：评论

学校读者我要写书评

暂无评论

Gain Scheduled PI controller design using Multi-Objective Reinforcement Learning

引用

IFAC-PapersOnLine 2024年第7期58卷 132-137页

作者： Kranthi Kumar P Ketan P Detroja Department of Artificial Intelligence Indian Institute of Technology Hyderabad India

关键词： Gain Scheduling Multi-Objective RL PI controller ppo algorithm Steam Turbine System Step response

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：