This repo contains the code for the paper "Reinforcement Learning for Dynamic Memory Allocation" - https://arxiv.org/abs/2410.15492

Video Summary - https://www.youtube.com/watch?v=m7EsBf1J5Pc

Installation

1. Clone the repository: git clone https://github.com/curry2736/rl-malloc.git
2. Install the required packages: pip install -r requirements.txt

Running the Code

Experiment 1

To run experiment 1 from the paper, execute the following command: python experiment1.py

If you would like to retrain from scratch, follow these steps:

1. Comment out line 24 in experiment1.py, which loads a previously trained model.
2. Uncomment lines 31 and 32 in experiment1.py for training.

Experiments 2 and 3

To run experiments 2 and 3 from the paper, open test_pol_sb3.ipynb.

To get the ith graph of experiment j:

1. Uncomment the 6 lines of code in the second cell which have the experiment i graph j commented above it.

Run the full notebook to get each individual graph