Implement additional modern algorithms

[barrybecker4]

The AI suggests:

Differential Evolution (DE) — Strong default for continuous bounded search; simple, few hyperparameters, parallel-friendly.

Evolution strategies / CMA-ES family — Sample efficiency and adaptation on many continuous problems; heavier implementation cost.

Particle Swarm Optimization (PSO) — Common baseline for continuous multimodal problems; overlaps somewhat with DE/ES choice.

Modern GA/EA machinery — Tournament selection, explicit crossover operators aligned with ParameterArray types, diversity maintenance—your GA is closer to “steady-state + heavy mutation” than textbook GA.

Iterated local search / variable neighborhood search — Fit discrete/permutation problems already modeled.

Surrogate / Bayesian optimization — Different cost model (expensive evaluations); optional add-on, not a drop-in for every Optimizee.

[barrybecker4]

Recommended to start:
Pick one of DE, PSO, or a simplified ES as a new OptimizationStrategyType, reusing ParameterArray neighbor/sampling hooks where possible. This gives users a post-2004 baseline without rewriting the library.