Skip to content

算法优化:新增Dueling DQN架构和优先经验回放#3273

Closed
joe-justin369 wants to merge 19 commits into
OpenHUTB:mainfrom
joe-justin369:main
Closed

算法优化:新增Dueling DQN架构和优先经验回放#3273
joe-justin369 wants to merge 19 commits into
OpenHUTB:mainfrom
joe-justin369:main

Conversation

@joe-justin369

Copy link
Copy Markdown
Contributor

修改概述:

修改的详细描述

  1. 算法架构升级:实现了Dueling DQN架构,将Q值分解为状态价值函数V(s)和动作优势函数A(s,a)
  2. 经验回放机制优化:实现了优先经验回放(PER)算法,根据TD误差优先级采样经验
  3. 模型结构增强:新增PrioritizedReplayBuffer类,支持优先级采样
  4. 监控和可视化改进:扩展训练统计,包含PER缓冲区使用情况

经过了什么样的测试?

  1. windows 11
  2. py版本3.7.9运行无报错,且符合预期
  3. Carla仿真流畅,符合代码优化预期

运行效果

动图、视频、截图等

2025-12-15.231232.mp4



MODEL_PATH = r'D:\Work\T_Unmanned_vehicle_AD_DQN\models\YY_Optimized_best_363.53.model' # 请替换为实际的最佳模型路径
MODEL_PATH = r'D:\Work\T_Unmanned_vehicle_AD_DQN\models\YY_Optimized_best_541.48.model' # 新模型路径

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

需要换成相对路径或者自动获取脚本所在的路径,否则其他机器不修改源代码跑不起来

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

我的下一个版本就新增了这个功能

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

image 我昨天也发现了

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants