Derringer-Suich desirability functions¶
In [1]:
import numpy as np
from moldrug import utils
import matplotlib.pyplot as plt
test_r = np.arange(0,3.25,0.25)
var1 = np.linspace(100, 500, 1000)
var2 = np.linspace(-15, 5, 1000)
var3 = np.linspace(0, 15, 1000)
# Vectorize the Desirability functions
f1 = np.vectorize(utils.LargerTheBest) # Here we are maximizing
f2 = np.vectorize(utils.SmallerTheBest) # Here we are minimizing
f3 = np.vectorize(utils.NominalTheBest) # Here we are looking for a range
In [2]:
fig, axs = plt.subplots(ncols=3, figsize = (30,9))
plt.setp(axs.flat, xlabel='Value', ylabel='desirably')
NUM_COLORS = len(test_r)
cm = plt.get_cmap('gist_rainbow')#gist_rainbow viridis
[ax.set_prop_cycle('color', [cm(1.*j/NUM_COLORS) for j in range(NUM_COLORS)]) for ax in axs]
for r in test_r:
d1 = f1(Value = var1, LowerLimit = 200, Target = 450, r = r)
d2 = f2(Value = var2, Target = -13.5, UpperLimit = 0, r = r)
d3 = f3(Value = var3, LowerLimit = 2, Target = 7, UpperLimit = 13, r1 = r, r2 = r)
axs[0].plot(var1, d1, label = str(r))
axs[1].plot(var2, d2, label = str(r))
axs[2].plot(var3, d3, label = str(r))
plt.legend()
axs[0].set(title = 'LargerTheBest')
axs[1].set(title = 'SmallerTheBest')
axs[0].set(title = 'NominalTheBest')
Out[2]:
[Text(0.5, 1.0, 'NominalTheBest')]
