# pendulardqn

**Repository Path**: DevilcLo/pendulardqn

## Basic Information

- **Project Name**: pendulardqn
- **Description**: 使用深度Q网络(Deep Q Network)学习控制单摆
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 1
- **Created**: 2025-05-07
- **Last Updated**: 2025-05-07

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# 使用深度Q网络(Deep Q Network)学习控制单摆

# C++测试代码
```cpp
#include <iostream>
#include "matplotlibcpp.h"
#include "env.hpp"
using namespace std;
namespace plt = matplotlibcpp;
int main(int argc, char **argv) {
    PendulumEnvironment env;
    vector<double> state;
    vector<double> ans1, ans2;
    int change = 30;
    if (argc >1) {
        change = atoi(argv[1]);
        cout << change << endl;
    }
    for (size_t i = 0; i < 300; i++)
    {
        if (i < change) env.Step(1);
        else env.Step(0);
        state = env.Get_State();
        ans1.push_back(state[0]);
        ans2.push_back(env.Get_Reward());
    }
    plt::plot(ans1);
    plt::plot(ans2);
    plt::show();
    return 0;
}
```

# 参考
[深度强化学习-DQN算法原理与代码 -CSDN博客](https://blog.csdn.net/weixin_46133643/article/details/121845874)  
译文 <https://blog.csdn.net/qq_34288751/article/details/130503825>
[__init__.py的神奇用法 -知乎](https://zhuanlan.zhihu.com/p/115350758)