Module evolution 

Regularized (aging) evolution for neural architecture search.

Implements the algorithm from:

Real et al. (2019) “Regularized Evolution for Image Classifier Architecture Search” https://arxiv.org/abs/1802.01548

The key idea is a cyclic population (VecDeque): the oldest individual is evicted whenever the population exceeds population_size, even if it happens to be the highest-scoring member. This aging pressure prevents premature convergence and keeps the search exploratory.

The ask/tell API lets the caller supply evaluation scores without any stored objective closure, matching the hyperparameter-optimization convention used elsewhere in this crate.

Structs

NasResult

Summary of a completed (or in-progress) NAS run.

RegularizedEvolution

Regularized (aging) evolution NAS searcher.