Pyomo - 将目标 function 放在索引上？

Question

I want to loop an objective function on an index. Let's say that I have the following dataframe:

index_i	index_j	value
10400	Mexico	23
10400	Guatemala	44
10600	Mexico	25
10600	Guatemala	30

I want to do something like this:

for i in model.i:
         model.obj = Objective(sum(x[i,j] for j in model.j), maximize)
         ...
         results = opt.solve(model)
 

I want to do the loop directly in the objective function or solve, because I don't want to build i-times the model variables and parameters (ie, looping the opt model i-times).

Thanks!!!

Answer 1

You have it correct in the first instance. It isn't clear what you are trying to avoid... ? Realize all this is doing is constructing an expression to hand to the model. It isn't "looping" as it solves, if that is your concern.

Your second effort, looping over i, will create a new obj for each value of i and then solve with that specific value only . Is that what you intend?

For example:

from pyomo.environ import *

m = ConcreteModel()

# SETS
m.I = Set(initialize=[1,2,3])
m.J = Set(initialize=list('ABC'))

# VARIABLES
m.x = Var(m.I, m.J, domain=Binary)

# GOOD OBJ
m.obj1 = Objective(expr=sum(m.x[i,j] for i in m.I for j in m.J))

# BAD OBJ!!  DON'T DO THIS
for i in m.I:
    m.obj2 = Objective(expr=sum(m.x[i,j] for j in m.J))


m.pprint()

Yields:

3 Set Declarations
    I : Size=1, Index=None, Ordered=Insertion
        Key  : Dimen : Domain : Size : Members
        None :     1 :    Any :    3 : {1, 2, 3}
    J : Size=1, Index=None, Ordered=Insertion
        Key  : Dimen : Domain : Size : Members
        None :     1 :    Any :    3 : {'A', 'B', 'C'}
    x_index : Size=1, Index=None, Ordered=True
        Key  : Dimen : Domain : Size : Members
        None :     2 :    I*J :    9 : {(1, 'A'), (1, 'B'), (1, 'C'), (2, 'A'), (2, 'B'), (2, 'C'), (3, 'A'), (3, 'B'), (3, 'C')}

1 Var Declarations
    x : Size=9, Index=x_index
        Key      : Lower : Value : Upper : Fixed : Stale : Domain
        (1, 'A') :     0 :  None :     1 : False :  True : Binary
        (1, 'B') :     0 :  None :     1 : False :  True : Binary
        (1, 'C') :     0 :  None :     1 : False :  True : Binary
        (2, 'A') :     0 :  None :     1 : False :  True : Binary
        (2, 'B') :     0 :  None :     1 : False :  True : Binary
        (2, 'C') :     0 :  None :     1 : False :  True : Binary
        (3, 'A') :     0 :  None :     1 : False :  True : Binary
        (3, 'B') :     0 :  None :     1 : False :  True : Binary
        (3, 'C') :     0 :  None :     1 : False :  True : Binary

2 Objective Declarations
    obj1 : Size=1, Index=None, Active=True
        Key  : Active : Sense    : Expression
        None :   True : minimize : x[1,A] + x[1,B] + x[1,C] + x[2,A] + x[2,B] + x[2,C] + x[3,A] + x[3,B] + x[3,C]
    obj2 : Size=1, Index=None, Active=True
        Key  : Active : Sense    : Expression
        None :   True : minimize : x[3,A] + x[3,B] + x[3,C]

6 Declarations: I J x_index x obj1 obj2

Note in the 2nd obj, it is overwriting itself in the loop and only captures the last value of i (3).

If this isn't what you were asking about, edit your question for clarity.