Skip to content

Architecture

EvoToolkit is designed with modularity and extensibility in mind.

Core Components

1. Tasks (evotoolkit.task)

Define optimization problems and evaluation logic.

2. Methods (evotoolkit.evo_method)

Implement evolutionary algorithms (EoH, EvoEngineer, FunSearch).

3. Interfaces (evotoolkit.core.method_interface)

Bridge between tasks and methods, handling algorithm-specific adaptations.

4. Registry (evotoolkit.registry)

Automatic discovery and registration of tasks and algorithms.

Design Patterns

  • Factory Pattern: evotoolkit.solve() creates algorithm instances
  • Strategy Pattern: Interfaces provide algorithm-specific strategies
  • Template Method: Base classes define workflow, subclasses customize

Module Organization

evotool/
├── core/               # Base classes and abstractions
├── evo_method/         # Algorithm implementations
├── task/               # Task implementations
├── tools/              # Utilities (LLM API, etc.)
└── registry.py         # Component registration

For detailed implementation guides, see the source code documentation.