Optimistix: modular optimisation in JAX and Equinox

2024/02/15 by Jason Rader, Terry Lyons, Rader, Jason +3 · 15 citations
Engineering · Materials Science · #Advanced Control Systems Optimization #FOS: Computer and information sciences #FOS: Mathematics #Innovative Microfluidic and Catalytic Techniques Innovation #Mathematical Software (cs.MS) #Optimization and Control (math.OC) #Synthesis and properties of polymers

paper · pdf · doi:10.48550/arxiv.2402.09983

openalex publication_date 2024/02/15 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We introduce Optimistix: a nonlinear optimisation library built in JAX and Equinox. Optimistix introduces a novel, modular approach for its minimisers and least-squares solvers. This modularity relies on new practical abstractions for optimisation which we call search and descent, and which generalise classical notions of line search, trust-region, and learning-rate algorithms. It provides high-level APIs and solvers for minimisation, nonlinear least-squares, root-finding, and fixed-point iteration. Optimistix is available at https://github.com/patrick-kidger/optimistix.

Cited by

• jaxdae: A JAX-native Differentiable Solver for Differential-Algebraic Equations in Coupled Multi-physics
• Thermo-Structural Topology Optimization Considering Nonlinear Creep
• ARCADE: Adaptive Robot Control with Online Changepoint-Aware Bayesian Dynamics Learning
• Autonomous Learning of Attractors for Neuromorphic Computing with Wien Bridge Oscillator Networks
• SoftJAX & SoftTorch: Empowering Automatic Differentiation Libraries with Informative Gradients
• Differentiable N-body code for Galactic Dynamics -- Odisseo
• Case study of a differentiable heterogeneous multiphysics solver for a nuclear fusion application
• A parallel-in-time Newton's method-based ODE solver
• Meshless solutions of PDE inverse problems on irregular geometries
• Illuminating the lantern: coherent, spectro-polarimetric characterisation of a multimode converter
• Automated design of compound lenses with discrete-continuous optimization
• linrax: A JAX Compatible, Simplex Method Linear Program Solver
• TOFLUX: A Differentiable Topology Optimization Framework for Multiphysics Fluidic Problems
• Towards real-time surrogate-free Bayesian inversion for neutron reflectometry
• ParamRF: A JAX-native Framework for Declarative Circuit Modelling

Related

• Lineax: unified linear solves and linear least-squares in JAX and Equinox
• optHIM: Hybrid Iterative Methods for Continuous Optimization in PyTorch
• OptiMUS-0.3: Using Large Language Models to Model and Solve Optimization Problems at Scale
• evosax: JAX-based Evolution Strategies
• Optuna: A Next-generation Hyperparameter Optimization Framework