cpagym

RL-Accounting 会计强化学习

Redefining Accounting! Developed by Bfsu Ai-Accounting Team (c) 2023 Github

Install & Import 安装和导入

For detailed installation instructions, see the documentation.

pip install cpagym
import gymnasium as gym
import cpagym as cg 
import torch

Accounting Envs 主要会计环境

1. Moral_Hazard 道德风险

Action Space

Num	Action	Unit
0	Discharge of duty	0
1	Due diligence	1

Observation Space

The observation is a ndarray with shape (6,) that provides information about the two rotational joint angles as well as their angular velocities:

Num	Observation	Min	Max
0	Project success rate	0	1
1	Project success return	0	1

env = gym.make("Moral_Hazard-v1")#Instantiate the Environment载入道德风险环境
state, info = env.reset(seed=2)#初始化
n_steps = 10
for _ in range(n_steps):
    action = env.action_space.sample()#随机行动
    state, reward, terminated, truncated, info = env.step(action)
    print(observation, reward, terminated, truncated, info )
    if terminated or truncated:
        observation, info = env.reset()

Accounting RL-Algorithm 会计强化学习算法

1. Train the Agent with DQN

Accounting RL-Agent 会计强化学习代理人

agent=cg.Agent(state_size=1,action_size=2)#Instantiate the Agent载入会计智能体
env = gym.make("Moral_Hazard-v1")
state, info = env.reset(seed=2)
action = agent.act(state, eps)#最优化行动
next_state, reward, done, _,_ = env.step(action)
agent.step(state, action, reward, next_state, done)
print(agent.memory.memory)

配套智能会计包CPANLP

Check out: https://cpanlp.com/reinforcement