Giter Site home page Giter Site logo

akahello / jobschedule Goto Github PK

View Code? Open in Web Editor NEW

This project forked from weixians/jobschedule

0.0 0.0 0.0 405 KB

Implementation of paper "Research on Adaptive Job Shop Scheduling Problems Based on Dueling Double DQN"

License: MIT License

Python 100.00%

jobschedule's Introduction

JobSchedule

Implementation of paper "Research on Adaptive Job Shop Scheduling Problems Based on Dueling Double DQN" (论文复现)

运行

  1. 依赖安装
pip3 install -r requirements.txt
  1. 训练
python3 run_ddqn.py --output=../output_jobshop --dueling 
  1. 测试
python3 run_ddqn.py --output=../output_jobshop --dueling --test
  1. 测试并生成图片
python3 run_ddqn.py --output=../output_jobshop --dueling --test --render --mode=img 

论文中一些解释含糊地方的个人解释

Observation

该observation由3个channel组成, 包含: Processing time, Scheduling finish, 和 Machine Utilization. 三者的行表示 job,列表示 operation (有的文章也叫 task). 结合图5和图11来看:

  1. Processing time: 初始化为每个operation所需的执行时长, 运行过程中, 取哪个operation出来, 就将位置的值置为0
  2. Scheduling finish: 该任务执行完成时的时刻, 赋值时需要判断该operation前一个完成时间和该operation对应机器完成时间哪个比较长,取长的那个
  3. Machine Utilization: 记录每个机器的利用率,注意这里为每个cell赋值时,值为该operation对应的机器的当前时间,而不是该operation之前的完成时刻

Reward

reward = 当前已完成所有operation的操作时间和/(机器数量*当前makespan)

一些疑问

  1. 第10页中指的已知最佳策略 $\pi^{best}$ 是什么?
  2. 第12页中 100*20 和 84*84 分别是什么意思?
  3. 训练时是不是只用一个instance (比如ft06,la16等) 一直训?
  4. 个人觉得,作者这里根据输入的大小会调整网络结构,不然比如ft06这个数据是无法使用图8所示的最佳网络结构的。虽然作者也从没给出任何一个详细完整的网络结构。

案例 la16 的训练过程展示

测试时最低makespan可得到1040左右 (似乎比论文中展示的效果还要好些)

jobschedule's People

Contributors

weixians avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.