An Analysis of the RBF Hyperparameter Impact on Surrogate-Assisted Evolutionary Optimization

Algorithm 2

The implemented RBF-based SAEA.

	Input: initial sample size, max evaluations, EA parameters, and parameters;
	generate and sample an initial random population;
	repeat
	train an RBF surrogate based on the evaluated vectors;
	run one generation of the EA;
	for eachgeneration do
	evaluate with the best vectors with the true function;
	retrain the surrogate and re-evaluate the population;
	end
	until max function evaluations or max EA generations;
	output: the best solution found;