OMMX Python SDK 1.5.0

This notebook describes the new features. Please refer the GitHub release note for the detailed information.

Evaluation and Partial Evaluation

From the first release of OMMX, ommx.v1.Instance supports evaluate method to produce Solution message

from ommx.v1 import Instance, DecisionVariable

# Create an instance of the OMMX API
x = DecisionVariable.binary(1)
y = DecisionVariable.binary(2)

instance = Instance.from_components(
    decision_variables=[x, y],
    objective=x + y,
    constraints=[x + y <= 1],
    sense=Instance.MINIMIZE
)
solution = instance.evaluate({1: 1, 2: 0})

solution.decision_variables

	kind	lower	upper	name	subscripts	description	substituted_value	value
id
1	binary	0.0	1.0	<NA>	[]	<NA>	<NA>	1.0
2	binary	0.0	1.0	<NA>	[]	<NA>	<NA>	0.0

From Python SDK 1.5.0, Function and its base classes, Linear, Quadratic, and Polynomial also support evaluate method:

f = 2*x + 3*y
value, used_ids = f.evaluate({1: 1, 2: 0})
print(f"{value=}, {used_ids=}")

value=2.0, used_ids={1, 2}

This returns evaluated value of the function and used decision variable IDs. If some decision variables are lacking, the evaluate method raises an exception:

try:
    f.evaluate({3: 1})
except RuntimeError as e:
    print(e)

Variable id (1) is not found in the solution

In addition, there is partial_evaluate method

f2, used_ids = f.partial_evaluate({1: 1})
print(f"{f2=}, {used_ids=}")

f2=Linear(3*x2 + 2), used_ids={1}

This creates a new function by substituting x = 1. partial_evaluate is also added to ommx.v1.Instance class:

new_instance = instance.partial_evaluate({1: 1})
new_instance.objective

Function(x2 + 1)

This method will be useful for creating a problem with fixing specific decision variables.