Summary of Offline Reinforcement Learning For Learning to Dispatch For Job Shop Scheduling, by Jesse Van Remmerden et al.
Offline Reinforcement Learning for Learning to Dispatch for Job Shop Scheduling
by Jesse van Remmerden, Zaharah Bukhsh, Yingqian Zhang
First submitted to arxiv on: 16 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper presents Offline Reinforcement Learning for Learning to Dispatch (Offline-LD), an approach that addresses limitations of online Reinforcement Learning (RL) in solving the Job Shop Scheduling Problem (JSSP). Online RL requires extensive training, cannot leverage existing solutions, and often yields suboptimal results. Offline-LD learns from previously generated solutions, motivated by scenarios where historical data is available. The method adapts CQL-based Q-learning methods for maskable action spaces, introduces an entropy bonus modification, and exploits reward normalization. Experiments show that Offline-LD outperforms online RL on generated and benchmark instances. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Offline-LD solves a big problem called the Job Shop Scheduling Problem (JSSP) with computers. This problem is like scheduling tasks for machines in a factory. Online computers can solve it, but they need to learn everything from scratch, which takes a long time. They also don’t do as well as other methods that already know some things about how to schedule tasks. Offline-LD is new way of doing things that learns from what’s already known. It works better than the online method and can even make good decisions with incomplete information. |
Keywords
* Artificial intelligence * Reinforcement learning
