About PyScat

PyScat is a Python library implementing scatter-search metaheuristics for global optimization.

Note

This package is still under development. Documentation is incomplete, and the API may change without prior notice.

Repository: https://github.com/ICB-DCM/pyscat

Installation

Releases from PyPI:

pip install pyscat

The latest development version from GitHub:

pip install git+https://github.com/ICB-DCM/pyscat.git#egg=pyscat

Algorithms

• Enhanced Scatter Search (eSS)

  Based on Egea et al. (2007)

• Self-Adaptive Cooperative enhanced Scatter Search (saCeSS)

  Based on Penas et al. (2017)

Contents:

• About
  • Installation
  • Algorithms
• Getting started
  • Setting up the optimization problem
  • Running the optimization
• eSS
  • What is scatter search?
  • Set up problem
  • Enhanced Scatter Search (eSS) — ESSOptimizer
    • Initialize ESS
    • Recombination — generate new points based on the current RefSet
    • Go-beyond
    • Local search (optional)
    • Replacement of stuck RefSet members
    • Full example
    • General hyperparameters for ESSOptimizer
• saCeSS
  • Set up problem
  • Self-Adaptive Cooperative Enhanced Scatter Search (saCeSS)—SacessOptimizer
    • Optimization with default options
    • saCeSS hyperparameters
    • Parallelization within SacessOptimizer
• How-to
  • How to create a parameter ensemble from an optimization run?
• API reference
  • pyscat
    • ESSExitFlag
    • ESSOptimizer
    • EvalLogger
    • SacessCmaFactory
    • SacessFidesFactory
    • SacessIpoptFactory
    • SacessOptimizer
    • SacessOptions
    • ThresholdSelector
    • TopKSelector
    • get_default_ess_options()
  • pyscat.plot
    • merge_monotonic_histories()
    • plot_ess_history()
    • plot_sacess_history()
• Release notes
  • pyscat-0.0.1 (2025-12-02)
    • Most relevant changes since pyPESTO 0.5.7
    • Migrating from pyPESTO 0.5.7 to PyScat 0.0.1
• Development
  • Development setup
  • Running tests
  • Building documentation