Bayesian optimization is a sequential design strategy for global optimization of black-box functions (such as fitness functions). Uses Gaussian processes as a prior. It implies diversity when selecting sequences (1):
“Once the acquisition function has been decided, it implicitly determines the appropriate notion of diversity in the resulting proposed batch of sequences.”
1.
Fannjiang C, Listgarten J. Is novelty predictable? arXiv. 2023; Available from: https://arxiv.org/abs/2306.00872