Back to evidence library

10 / Educational simulation · 2026

Public non-clinical simulationdata sciencequant

Adaptive dosing as an educational RL simulation

A synthetic control benchmark with a strict non-clinical boundary.

The work is deliberately framed as teaching software. Its synthetic parameters, reward and dynamics cannot support patient dosing or clinical claims.

Adaptive Dosing project visual

Evidence register

20×20

state grid

Explicit tabular representation

4/4

tests

Environment and API contracts

held-out

seeds

Separate from training episodes

01 / Problem

An interesting controller can become harmful if a synthetic experiment is presented as a medical recommendation.

02 / Approach

Define a small MDP, compare learned and fixed synthetic policies on unseen seeds, and enforce the educational boundary in API and documentation.

03 / Outcome

A coherent end-to-end RL demonstration now includes evaluation, tests, CI, Docker and a strict simulation interface.

How the evidence is produced.

Repositioned and rebuilt the concept as bounded educational software with reproducible evaluation.

  1. 01Synthetic ODE → discrete state
  2. 02Q-learning → policy
  3. 03Held-out seeds → comparison
  4. 04Strict API → simulation only

Validation scope

Four tests pass; evaluation compares policies on held-out random seeds under the same synthetic environment.

Known limitation

No patient data, pharmacological calibration, clinical utility or safety validation. It must never recommend a real dose.

What is inspectable

  • Non-clinical language is enforced.
  • Training and evaluation seeds are separated.
  • A fixed policy remains visible as baseline.

Next proof to add

  1. 01Add sensitivity analysis over dynamics.
  2. 02Report uncertainty across many seeds.
  3. 03Keep all outputs explicitly synthetic.

Main stack

PythonQ-learningODEFastAPIDockerCI