ReplenishmentRL

Replenishment environment for OR and RL algorithms

Contents

Folder	Description
ReplenishmentEnv	Replenishment env source code.
ReplenishmentEnv\config	Config for building the env.
ReplenishmentEnv\data	Csv based data for skus, including sku_list, info and other dynamic data
ReplenishmentEnv\env	Kernel simulator for env
ReplenishmentEnv\utility	Utility simulator for env and wrapper
ReplenishmentEnv\wrapper	Wrapper for env(Not implement yet)c
Example	Show case of replenishment env.

Install from Source

• Download code

git clone https://github.com/zhangchuheng123/ReplenishmentRL.git

• Build and install

python setup.py install

• Run example

import random
from ReplenishmentEnv import make_env

env = make_env("sku58")
env.reset()
for i in range(10):
    action_list = [int(random.random() * 10) for i in range(58)]
    states, rewards, done, infos = env.step(action_list) 
print(infos["balance"])

Run example without installation

• Download code

git clone https://github.com/zhangchuheng123/ReplenishmentRL.git

• Build the environment

  Only python >= 3.8.0 is tested.

  • Build by conda

  conda create -n ReplenishmentEnv python=3.8.0 -f requirements.txt
  

  • Build by pip

  pip install -r requirements.txt
  

• Run example

    python Example\random_actions.py

Build new env(Not Implement Yes)

• Prepare the data including sku_list, info and dynamic data.
• Prepare the env config following the instruction in Demo.yml
• Prepare the special or current wrapper
• Register in ReplenishmentEnv\init.py file