pc‑gym Dataset Overview

Dataset Introduction

pc‑gym is a collection of environments for process‑control reinforcement learning. It offers several simulated environments for training and testing RL algorithms in process‑control applications.

Quick Start

Below is a simple example showing how to set up a continuous stirred‑tank reactor (CSTR) environment and run a simulation:

import pcgym

# Simulation parameters
nsteps = 100
T = 25

# Setpoint
SP = {Ca: [0.85 for i in range(int(nsteps/2))] + [0.9 for i in range(int(nsteps/2))]}

# Action and observation spaces
action_space = {low: np.array([295]), high: np.array([302])}
observation_space = {low: np.array([0.7,300,0.8]),high: np.array([1,350,0.9])}

# Build environment parameter dictionary
env_params = {
    N: nsteps, # time steps
    tsim:T, # simulation time
    SP :SP, 
    o_space : observation_space, 
    a_space : action_space, 
    x0: np.array([0.8, 330, 0.8]), # initial conditions [Ca, T, Ca_SP]
    model: cstr_ode, # select model
}

# Create environment
env = pcgym.make_env(env_params)

# Reset environment
obs, state = env.reset()

# Sample a random action
action = env.action_space.sample()

# Step in the environment
obs, rew, done, term, info = env.step(action)

Documentation

Full documentation can be found here.

Installation

Install the latest version via PyPI:

pip install pcgym

Examples

Example notebooks covering training loops, constraint handling, disturbances and policy evaluation are available here.

Implemented Process‑Control Environments

Citation

If you use pc‑gym in your research, please cite:

@software{pcgym2024,
  author = {Max Bloor and  Jose Neto and Ilya Sandoval and Max Mowbray and Akhil Ahmed and Mehmet Mercangoz and Calvin Tsay and Antonio Del Rio‑Chanona},
  title = {{pc‑gym}: Reinforcement Learning Environments for Process Control},
  url = {https://github.com/MaximilianB2/pc-gym},
  version = {0.1.6},
  year = {2024},
}