Examples

example/ holds runnable algorithms and demos. It is excluded from the built wheel. Every optimizer solves a reference test function from opfunu-core, and engineering_demo.py solves a constrained problem from enoppy-core — install both with:

uv sync --extra benchmarks

Optimizer implementations are grouped by algorithm family under example/optimizers/; files that demonstrate the API itself stay at the top level.

Getting started

File	Shows
hello_world_v1.py	@cm.declare function style
hello_world_v2.py	@cm.template class style + @cm.vectorize / @cm.precompile kernels
operator_module.py	@cm.module Operator Modules + cm.Constant, reused across two models
reduce_demo.py	@cm.reduce with annotation-inferred layout
engineering_demo.py	constrained engineering problem via enoppy-core

Optimizers

Each directory under example/optimizers/ collects the implementations of one well-known optimizer.

de/ — Differential Evolution

File	Algorithm
de/de_algorithm.py	Differential Evolution (jDE-style self-adaptive)
de/jade_algorithm.py	JADE with external archive
de/lshade_algorithm.py	LSHADE with success-history adaptation + linear population-size reduction

gwo/ — Grey Wolf Optimizer

File	Algorithm
gwo/gwo_algorithm.py	Grey Wolf Optimizer (pure Python reference)
gwo/gwo_algorithm_vectorized.py	GWO with numba kernels (GreyWolfOptimizer(parallel=True))

One-file families

Directory	File	Algorithm
pso/	pso/pso_algorithm.py	Particle Swarm Optimization
ga/	ga/ga_algorithm.py	Real-coded Genetic Algorithm (self-adaptive, elitist)
sa/	sa/sa_algorithm.py	Simulated Annealing as a population of chains
rs/	rs/random_search.py	Adaptive Random Search (Luus–Jaakola style)

Every one of them solves an opfunu-core function through the f.evaluate / f.bounds pattern described in Test functions.

Run any of them from the repo root, e.g.:

.venv/bin/python example/reduce_demo.py
.venv/bin/python example/hello_world_v1.py
.venv/bin/python example/optimizers/de/jade_algorithm.py