Very small (~e-32) expected improvement predictions for maximized objective and increasing expected improvements for minimized objective

[angiorda]

I am working with a scope of 24 experiments (batch = 3, rounds=8) with an objective to max (Intensity) and an objective to minimize (CoV). I have noticed that for the expected improvement of the Intensity objective gets smaller each round (which is good) but extremely small (max ~1e-32) and the expected improvement for the coefficient of variation gets bigger each round. I have included the max per round EI numbers, the max EI graphs for each objective, the hypervolume graph and Pareto front graph. Additionally, I included the .csv files. Any ideas on what I am doing wrong?

pred_spin_coating_opt_scope_R1.csv
pred_spin_coating_opt_scope_R2.csv
pred_spin_coating_opt_scope_R3.csv
pred_spin_coating_opt_scope_R4.csv
pred_spin_coating_opt_scope_R5.csv
pred_spin_coating_opt_scope_R6.csv
pred_spin_coating_opt_scope_R7.csv