def test_param_from_pandas(self):
# Test issue #68
model = ConcreteModel()
model.I = Set(initialize=range(6))
model.P0 = Param(model.I,
initialize={
0: 400.0,
1: 0.0,
2: 0.0,
3: 0.0,
4: 0.0,
5: 240.0
})
model.P1 = Param(model.I,
initialize=pd.Series({
0: 400.0,
1: 0.0,
2: 0.0,
3: 0.0,
4: 0.0,
5: 240.0
}).to_dict())
model.P2 = Param(model.I,
initialize=pd.Series({
0: 400.0,
1: 0.0,
2: 0.0,
3: 0.0,
4: 0.0,
5: 240.0
}))
#model.pprint()
self.assertEqual(list(model.P0.values()), list(model.P1.values()))
self.assertEqual(list(model.P0.values()), list(model.P2.values()))
model.V = Var(model.I, initialize=0)
def rule(m, l):
return -m.P0[l] <= m.V[l]
model.Constraint0 = Constraint(model.I, rule=rule)
def rule(m, l):
return -m.P1[l] <= m.V[l]
model.Constraint1 = Constraint(model.I, rule=rule)
def rule(m, l):
return -m.P2[l] <= m.V[l]
model.Constraint2 = Constraint(model.I, rule=rule)
def __init__(self, data, weights=None, nvar=None, objective=True):
if nvar is None:
nvar = len(data)
if weights is None:
weights = np.ones(len(data)) / nvar
if len(data) != len(weights):
print("Data and weights don't match in length")
return
model = ConcreteModel()
# Number of original variables
model.nvar = Param(within=NonNegativeIntegers, initialize=nvar)
# Number of lines in data
model.n = Param(within=NonNegativeIntegers,
initialize=np.shape(data)[0])
# Number of columns in data
model.m = Param(within=NonNegativeIntegers,
initialize=np.shape(data)[1])
model.I = RangeSet(0, model.n - 1)
model.J = RangeSet(0, model.m - 1)
# Initialize all x_ij = 0.0, when j != 0, and all x_i0 = 1.0
model.x = Var(model.I, model.J, domain=Binary, initialize=0)
for i in model.I:
model.x[i, 0].value = 1.0
# Initialize c_ij from given data
def c_init(model, i, j):
return data[i, j]
model.c = Param(model.I, model.J, initialize=c_init)
# Initialize w_i from given parameter)
def w_init(model, i):
return weights[i]
model.w = Param(model.I, initialize=w_init)
if objective:
model.OBJ = Objective(rule=self.obj_fun, sense=maximize)
def const(model, i):
''' Constraint: Given line i has only one 1
\sum_{i=1}^{n}x_{ij} = 1'''
return sum(model.x[i, j] for j in model.J) == 1
model.Constraint1 = Constraint(model.I, rule=const)
self.model = model
self._modelled = True
def __init__(self, data):
self._solved = False
self.res = None
model = ConcreteModel()
# Number of lines in data
model.n = Param(within=NonNegativeIntegers, initialize=len(data))
# Number of columns in data
model.m = Param(within=NonNegativeIntegers, initialize=len(list(data)))
model.I = RangeSet(0, model.n - 1)
model.J = RangeSet(0, model.m - 1)
# Initialize all x_ij = 0.0, when j != 0, and all x_i0 = 1.0
model.x = Var(model.I, model.J, domain=Binary, initialize=0.0)
for i in model.I:
model.x[i, 0].value = 1.0
# Initialize c_ij from given data
def c_init(model, i, j):
return data.values[i, j]
model.c = Param(model.I, model.J, initialize=c_init)
'''Objective function: Formulate problem as binary problem.
\sum_{i=1}^{n} \sum_{j=1}^{m} c_{ij}*x_{ij}'''
def obj_fun(model):
return sum(
sum(model.x[i, j] * model.c[i, j] for i in model.I)
for j in model.J)
model.OBJ = Objective(rule=obj_fun, sense=maximize)
''' Constraint: Given line i has only one 1
\sum_{i=1}^{n}x_{ij} = 1'''
def const(model, i):
return sum(model.x[i, j] for j in model.J) == 1
model.Constraint1 = Constraint(model.I, rule=const)
self.model = model
self._modelled = True
model.OBJ = Objective(expr=obj_expression(model))
def _e(model):
return np.dot(values, list(model.ct))
model.e = Expression([0], rule=_e)
def constr(model):
return model.e == 88
model.Constraint1 = Constraint(expr=constr(model))
def ObjRule(model):
return 2 * model.x[1] + 3 * model.x[2]
model.g = Objective(rule=ObjRule)
prob = model.create()
optim = SolverFactory('glpk')
result = optim.solve(prob, tee=True)
prob.load(result)
# all variables
prob.display()
