optimatika/ojAlgo

Pure Java LP, QP and MIP solvers plus fast linear algebra. Zero dependencies, MIT licensed. Mathematical optimization and linear algebra for the JVM.

What it solves

ojAlgo provides a high-performance, zero-dependency Java library for mathematical operations, specifically targeting linear algebra and mathematical programming (optimisation). It allows developers to solve complex numerical problems without relying on external native libraries.

How it works

The library is built on a flexible system of array classes that support sparse or dense formats and can handle various number types (including complex, rational, and quaternions). It uses these arrays to implement:

  • Linear Algebra: High-speed matrix operations.
  • Optimisation Solvers: Tools for Linear Programming (LP), Quadratic Programming (QP), and Mixed-Integer Programming (MIP).
  • Data Science Utilities: Implementations of Artificial Neural Networks and clustering algorithms.

Who it’s for

Java developers who need professional-grade mathematical tools for data science, financial modelling, or industrial optimisation problems (such as bin packing or shift scheduling) within a pure Java environment.

Highlights

  • Zero Dependencies: Pure Java implementation requiring no external libraries.
  • High Performance: Ranked as one of the fastest pure Java linear algebra libraries in independent benchmarks.
  • Flexible Memory: Arrays can be allocated off-heap or stored in a file to handle arbitrarily large datasets.
  • Broad Mathematical Scope: Covers everything from stochastic processes and descriptive statistics to complex number handling.

Related

  • Project
    knuddelsgmbh/jtokkit

    JTokkit is a zero‑dependency Java library that reproduces OpenAI’s tokenisation algorithms (e.g., cl100k_base) for fast, accurate token counting and conversion on the JVM. It offers a thread‑safe API, Maven/Gradle distribution, and extensibility for custom encodings, all under an MIT license.

  • Project
    oracle/tribuo

    A Java machine learning library that provides a unified interface for classification, regression, clustering, and anomaly detection, featuring built-in model provenance and ONNX export.

  • Project
    dromara/Omega-AI

    A deep learning framework built in Java that supports a wide range of neural network architectures from CNNs to LLMs and Diffusion models with CUDA acceleration.

  • Project
    google-deepmind/simply

    A minimal and scalable JAX-based codebase for frontier LLM research designed for rapid iteration by both human researchers and AI agents.

  • Project
    haifengl/smile

    A comprehensive, high-performance machine learning framework for the JVM that integrates traditional statistical ML, deep learning, and LLM inference.