Table 1.

List of hyperparameters for CACLA and MLP subject to evolutionary search.

Parameter name	Symbol	Search space
CACLA + var beta	β	{k : k + 0.00025, 0.0001 ≤ k ≤ 0.0251}
Initial variance	var0	{k : k + 0.1, 2.0 ≤k ≤7.5}
Initial iterations	$⌈{\mathrm{\delta }}_0∕\sqrt{\mathit{va}{r}_0}⌉$	{k : k + 1, 5 ≤ k ≤15}
Discount factor	γ	{k : k + 0.001, 0.60 ≤ k ≤1.0}
Exploration rate	σ	{k : k + 0.1, 0.7 ≤ k ≤1.6}
Exploration rate decay	κ	{k : k + 0.001, 0.940 ≤ k ≤1.0}
Learning rate Critic	η	{k : k + 0.0025, 0.0001 ≤ k ≤0.175}
Critic MLP in →h1st	Cin→h1st	{k : k + 5, 10 ≤ k ≤50}
Critic MLP h2nd →out	Ch2nd→out	{k : k + 5, 0 ≤ k ≤25}
Learning rate Actor	α	{k : k + 0.0025, 0.0001 ≤ k ≤0.175}
Actor MLP in →h1st	Ain→h1st	{k : k + 5, 10 ≤ k ≤50}
Actor MLP h2nd →out	Ah2nd→out	{k : k + 5, 0 ≤ k ≤25}
Reward	R	{1, 10, 100}
Punishment	P	{−1, −0.1, 0}

Variables range based on Navarro-Guerrero (2016).