Coverage for robust/synthetic_model/models.py : 0%

from __future__ import division 

from builtins import range 

import numpy as np 

from gpkit import Variable, Model 

def synthetic_model(number_of_constraints): 

constraints = [] 

obj = 1 

number_of_gp_variables = int(number_of_constraints/2) + int(number_of_constraints*np.random.rand()) + 1 

gp_variables = [] 

s = [] # Variable('s_relax_sm') 

for i in range(number_of_gp_variables): 

x = Variable('x_sm_%s' % i) 

gp_variables.append(x) 

constraints.append(x >= 0.01) 

number_of_uncertain_variables = int(50*np.random.rand()) + 1 

uncertain_variables = [] 

for i in range(number_of_uncertain_variables): 

uncertain_variables.append(Variable('u_sm_%s' % i, 2*np.random.random(), pr=50*np.random.random())) 

for counter in range(number_of_constraints): 

number_of_monomials = int(15*np.random.random())+1 

vector_to_choose_from = [0, 0, 0, 1, -1] # , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 

m = number_of_monomials*[1] 

for j in range(number_of_monomials): 

for _ in range(int(number_of_gp_variables*np.random.rand()/2) + 1): 

m[j] *= (np.random.choice(gp_variables))**(10*np.random.random()) # -5) 

for i in range(number_of_uncertain_variables): 

neg_pos_neutral_powers = [vector_to_choose_from[int(len(vector_to_choose_from)*np.random.rand())] for _ in range(number_of_monomials)] 

for j in range(number_of_monomials): 

m[j] *= uncertain_variables[i]**(np.random.rand()*2*(neg_pos_neutral_powers[j])) 

s.append(Variable('s_relax_sm_%s' % counter)) 

constraints.append(sum(m) <= s[counter]) 

for x in gp_variables: 

obj += 1000*np.random.rand()*x**(-np.random.rand()*10) # - 5) 

obj += sum([i**0.2 for i in s]) 

m = Model(obj, constraints) 

return m 

def test_synthetic_model(): 

a = Variable("a", 1.17, "-", "form factor", pr=10) 

x = Variable('x') 

constraints = [] 

constraints += [a*x + x/a + x**2/a**2 + a**3*x**1.3<= 1] 

return Model(1/x, constraints)