# pnnl/neuromancer

Pytorch-based framework for solving parametric constrained optimization problems, physics-informed system identification, and parametric model predictive control.

Repository: https://github.com/pnnl/neuromancer
Canonical: https://ross.abutalabs.com/products/neuromancer
Homepage: https://pnnl.github.io/neuromancer/
Language: Python
License: NOASSERTION
License Family: other
Topics: constrained-optimization, control-systems, deep-learning, differentiable-programming, dynamical-systems, pytorch, differentiable-optimization, nonlinear-dynamics, nonlinear-optimization, differentiable-control, physics-informed-ml
Last push: 2026-08-05T16:33:50+00:00

## Health v2 (maintenance only)
Score: 72/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 96, release rhythm 25, longevity 100
- inputs: {"age_days": 2149, "days_push": 28, "days_rel": 341, "gap_med": 98.0, "n_releases_24m": 5}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1369, forks 188 (observed 2026-08-28T04:04:31.835053+00:00)

## What it is
NeuroMANCER is a PyTorch-based differentiable programming library for solving parametric constrained optimization problems, physics-informed system identification, and parametric model-based optimal control. It provides a symbolic programming interface for embedding physics equations, domain knowledge, and constraints into end-to-end differentiable machine learning models.

## Use cases
- solve parametric constrained optimization problems with neural networks
- perform physics-informed system identification of dynamical systems
- learn neural controllers via differentiable predictive control
- model fluid dynamics with physics-informed neural networks
- train differentiable optimization layers with safety constraints
- learn Koopman operators or SINDy models of nonlinear dynamics
- optimize building HVAC control with learned models

## When to choose
- you need differentiable constrained optimization embedded in PyTorch training loops
- you want to combine scientific computing and machine learning with physics priors
- you are researching learning-to-optimize, learning-to-model, or learning-to-control methods
- you need parametric model predictive control for nonlinear systems

## When to avoid
- you need a general-purpose deep learning framework without optimization or control focus
- you require guaranteed optimal solutions to hard constrained problems rather than learned approximations
- you work outside Python/PyTorch ecosystems

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, simulation, math
- domain: machine-learning, deep-learning
- platform: python, cross-platform
- tags: differentiable-programming, constrained-optimization, model-predictive-control, physics-informed-neural-networks, system-identification, pytorch, scientific-machine-learning, control-systems, algorithms, automation

## Member repositories
- pnnl/neuromancer (main) score 72

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:31.835053+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:41:00.682097+00:00, confidence not recorded.
  - readme: https://github.com/pnnl/neuromancer (fetched 2026-08-28T04:04:31.835053+00:00, sha e65c4e9dcaf4)
  - homepage: https://pnnl.github.io/neuromancer/ (fetched 2026-08-29T11:57:55.332485+00:00, sha f53e7ea59543)
  - registry_pypi: https://pypi.org/pypi/neuromancer/json (fetched 2026-08-29T11:57:55.341753+00:00, sha 685be53184d1)
- Data as of 2026-08-30T08:39:29.467469+00:00.