heartpole 1.2

A simple Healthcare-themed Gymnasium for benchmarking Reinforcement Learning algorithms

HeartPole

The Why

HeartPole is a simple Gymnasium for benchmarking Reinforcement Learning algorithms. Think CartPole for healthcare. It is meant as a first test for reinforcement learning techniques aimed at the Healthcare domain, before moving on to actual clinical data like MIMIC III

This project includes a tutorial on creating your own Gymnasium environment, using HeartPole as an example, as well as a writeup of experimental results, submitted to HEALTHINF conference.

The What

You are a developer working on an important project. You want to maximize your productivity, but not at the expense of health - any serious health issue will negate all the effects of increased productivity.

Action space is discrete with 4 possible actions:

[do_nothing, drink_coffee, drink_beer, sleep]

Observation space is continious, with 6 dimensions:

['alertness', 'hypertension', 'intoxication',
 'time_since_slept', 'time_elapsed', 'work_done']

Your productivity mainly depends on alertness, you get a reward of 1 for every unit of work done and a reward of -100 if you suffer a heart attack.

The How

Install with

pip install heartpole

Create an environment with

from heartpole import HeartPole
env = HeartPole()

and use your favourite reinforcement learning algorithm, for example, from keras-rl