2

我正在尝试使用 PyOMO 解决约束混合整数非线性优化问题。具体来说，我试图找到满足两个给定齿轮比的齿轮直径和齿数。我对如何使用Set()和Var(). 我一直在阅读文档，但对于 Set 实际上是什么并不是超级清楚！它是我可以用来访问问题的类似分组部分的索引吗？这是我的代码：（Python 3.5）

from pyomo.environ import *
from pyomo.opt import SolverFactory
import numpy as np

# Define Forward and Reverse Gear Ratios

fwd_ratio = 4.3
rev_ratio = 9.1

D_guess = [4.5, 11.5, 6.0, 10.0, 4.5, 2.5, 2.25, 9.0]
N_guess = [18, 46, 24, 40, 18, 20, 18, 72]

idx = np.arange(0,8)

print(idx)

model = AbstractModel()

# Declare Model Sets??? I tried this as first argument to Var(), didn't work

#model.Didx = Set(D_guess)
#model.Nidx = Set(N_guess)

# Declare Model Variables

model.D = Var(D_guess, within='PositiveReals', bounds=(1.0,None))
model.N = Var(N_guess, within='PositiveInteger', bounds=(18,None))

# Declare Objective Functions

def obj_funcD(model):

    F1 = (model.D[1]/model.D[0])*(model.D[3]/model.D[2]) - fwd_ratio

    F2 = (model.D[1]/model.D[4])*(model.D[6]/model.D[5])*(model.D[7]/model.D[6]) - rev_ratio

    return F1 + F2

def obj_funcN(model):

    F1 = (model.N[1]/model.N[0])*(model.N[3]/model.N[2]) - fwd_ratio

    F2 = (model.N[1]/model.N[4])*(model.N[6]/model.N[5])*(model.N[7]/model.N[6]) - rev_ratio

    return F1 + F2

# Declare Constraint

def con_func1(model):

    return model.D[1]/model.D[0] == model.N[1]/model.N[0]

def con_func2(model):

    return model.D[3]/model.D[2] == model.N[3]/model.N[3]

def con_func3(model):

    return model.D[1]/model.D[4] == model.N[1]/model.N[4]

def con_func4(model):

    return model.D[6]/model.D[5] == model.N[6]/model.N[5]

def con_func5(model):

    return model.D[7]/model.D[6] == model.N[7]/model.N[6]

# Create Constraint List

model.c1 = Constraint(rule=con_func1)
model.c2 = Constraint(rule=con_func2)
model.c3 = Constraint(rule=con_func3)
model.c4 = Constraint(rule=con_func4)
model.c5 = Constraint(rule=con_func5)

# Create Objectives

model.obj1 = Objective(rule=obj_funcD,sense='minimize')
model.obj2 = Objective(rule=obj_funcN,sense='minimize')

# Solve the Problem?

opt = SolverFactory('glpk')

instance = model.create_instance()

results = opt.solve(instance)

此代码给出以下错误：

WARNING: Element 4.5 already exists in set D_index; no action taken.
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/pyomo/core/base/PyomoModel.py", line 920, in _initialize_component
ERROR: Constructing component 'D' from data=None failed:
    declaration.construct(data)
    ValueError: PositiveReals is not a valid domain. Variable domains must be an instance of one of (<class 'pyomo.core.base.set_types.RealSet' at 0x1004bee98>, <class 'pyomo.core.base.set_types.IntegerSet' at 0x1004f2558>, <class 'pyomo.core.base.set_types.BooleanSet' at 0x1004f28f8>), or an object that declares a method for bounds (like a Pyomo Set). Examples: NonNegativeReals, Integers, Binary
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/pyomo/core/base/var.py", line 573, in construct
    component=None)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/pyomo/core/base/var.py", line 299, in __init__
    "Integers, Binary" % (domain, (RealSet, IntegerSet, BooleanSet)))
ValueError: PositiveReals is not a valid d

I've also tried using RangeSet() and passing the associated Set as the first argument of Var() but this doesn't do anything either! I known I'm missing something super obvious here but I've been staring at the screen for 4 hours now and I'm enlisting your help! Thanks

4

1 回答 1

1

Change within='PositiveReals' to within=PositiveReals, etc.

The within (or domain) keyword should be assigned one the of set domain objects that get imported from pyomo.environ. They should not be assigned strings.

于 2017-04-26T02:23:11.420 回答