scikit opt

These evolutionary and swarm-based algorithms are useful for optimization problems where gradient information is unavailable, the objective function is non-convex with many local optima, or the search space is discrete or combinatorial.

The library is designed for practical usability each algorithm accepts a callable objective function and a parameter bounds specification, handles population initialization and evolutionary operators internally, and returns the best solution found along with convergence history for analysis.

Scikit-Opt includes implementations tuned for classic benchmark problems (traveling salesman, knapsack, function optimization) as well as utilities for visualizing optimization trajectories and comparing algorithm performance across problem instances.

Scikit-Opt is open-source under the MIT license and is used in engineering design optimization, hyperparameter tuning, feature selection, scheduling problems, and scientific research where classical gradient-based optimization is inapplicable.

The library is particularly accessible for researchers and engineers with Python backgrounds who need population-based optimization without implementing algorithm internals from scratch.

It is installable via pip and requires only NumPy and matplotlib as dependencies, making it lightweight relative to comprehensive optimization suites like DEAP or PyGMO.