AutoOptLib documentation

AutoOptLib automatically designs component-based metaheuristic optimizers on training instances and evaluates the selected designs on held-out instances.

User guide

• Getting started
  • Installation
  • Design and held-out evaluation
• Custom problems
  • Static problems
  • Evaluation budgets
  • Advanced definitions
• Application problems
  • Raw-material stacking
  • RIS passive beamforming
• Reproducibility
• Reliable objective evaluation
  • Checkpoint and resume
• Public API
  • autoopt(**options)
  • autooptlib.aldes
  • make_problem(...)
  • save_algorithm(design, path, metadata=None)
  • load_algorithm(path)
  • register_component(...)
  • Errors and constraints
• Architecture

Project

• Project information