How can you appropriately index a pyomo parameter that is initialized with a list by a multi-dimensional pyomo set? - pyomo

Question

For my application, I am trying to initialize a parameter for a concrete model using a function rule that outputs a list. The pyomo set I am using to index this parameter is multi-dimensional. It is important for this set to be multi-dimensional because it makes accessing the data structure I am pulling values from for the parameter much simpler. However, when I attempt to index with the set this way, I receive an index error.
Here is a simple test code that illustrates my issue.
First, I have my imports as necessary
## Testing pyomo sets and indexing
import pyomo.environ as pyo
import numpy as np
import itertools as iter
Then I define a concrete model and some numpy arrays
M = pyo.ConcreteModel()
a = np.arange(3)
b = np.arange(3)
Using an itertools product, I generate a list of two-dimensional tuples
T = list(iter.product(a,b))
M.T = pyo.Set(initialize = T)
M.T.pprint()
The output of this print is
T : Size=1, Index=None, Ordered=Insertion
Key : Dimen : Domain : Size : Members
None : 2 : Any : 9 : {(0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2)}
I define a function that returns a list and define a pyomo parameter that initializes with this function.
def paramListInitialize(M):
List = []
for i in M.T:
List.append(i)
return List
M.Param3 = pyo.Param(M.T, initialize = paramListInitialize(M))
I receive the following error.
ERROR: Rule failed for Param 'Param3' with index 0: KeyError: "Index '0' is
not valid for indexed component 'Param3'"
ERROR: Constructing component 'Param3' from data=None failed: KeyError: "Index
'0' is not valid for indexed component 'Param3'"
KeyError: "Index '0' is not valid for indexed component 'Param3'"
I am confused because I am able to define pyomo variable with this multi-dimensional index set and I can initialize parameters with lists if the index set in one dimensional; however, I am unable to get the pyomo object to associate the tuples in the pyomo set to the values in the initialization list.
It would be very helpful to know why this is not working for me or if there is another way to generate a multi-dimensional pyomo set that is not in the form of a list of tuples.

Answer 1

You need to look more closely at the dox on how to use a rule to initialize here.
The rule needs to return a single value for the index or indices passed in.
Here is an example using your 2-dim set:
import pyomo.environ as pyo
import numpy as np
import itertools as iter
m = pyo.ConcreteModel()
a = np.arange(3)
b = np.arange(3)
T = list(iter.product(a,b))
m.T = pyo.Set(initialize = T)
def example(m, a, b):
return a + b
m.p = pyo.Param(m.T, initialize=example)
m.pprint()
Yields:
1 Set Declarations
T : Size=1, Index=None, Ordered=Insertion
Key : Dimen : Domain : Size : Members
None : 2 : Any : 9 : {(0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2)}
1 Param Declarations
p : Size=9, Index=T, Domain=Any, Default=None, Mutable=False
Key : Value
(0, 0) : 0
(0, 1) : 1
(0, 2) : 2
(1, 0) : 1
(1, 1) : 2
(1, 2) : 3
(2, 0) : 2
(2, 1) : 3
(2, 2) : 4
2 Declarations: T p