nephrosim 0.1.0

Gymnasium environments for reinforcement learning in hemodialysis optimization

NephroSim

Gymnasium environments for reinforcement learning in hemodialysis optimization.

Installation

pip install nephrosim

Quick Start

import gymnasium as gym
import nephrosim

env = gym.make("nephrosim/UreaClearing-v0")
obs, info = env.reset()

for _ in range(240):
    action = env.action_space.sample()
    obs, reward, terminated, truncated, info = env.step(action)
    if terminated or truncated:
        break

Environments

Environment	Observation	Action	Objective
UreaClearing-v0	7-dim	2-dim (Qb, Qd)	Maximize Kt/V
UltrafiltrationControl-v0	8-dim	1-dim (UF rate)	Remove fluid safely
PhosphateMgmt-v0	6-dim	1-dim (binder dose)	Weekly phosphate control
FullDialysisSession-v0	11-dim	3-dim (Qb, Qd, UF)	Multi-objective

Citation

@article{dhia2026nephrosim,
  title={NephroSim: Gymnasium Environments for Reinforcement Learning in Hemodialysis Optimization},
  author={Dhia, Hass},
  year={2026},
  institution={Smart Technology Investments Research Institute}
}

License

MIT