Common Mistakes

Here are some common mistakes that users make.

Forgetting Adding a Transition

When you define a transition in the model, do not forget to add it to the model using add_transition().

import didppy as dp

model = dp.Model()
var = model.add_int_var(target=0)

transition = dp.Transition(
    name='increment',
    cost=1 + dp.IntExpr.state_cost(),
    effects=[(var, var + 1)],
)
# Do not forget this!
model.add_transition(transition)

Using Built-in max() for Expressions

When you want to take the maximum of two IntExpr, you cannot use built-in max() in Python. Instead, you need to use didppy.max().

import didppy as dp

model = dp.Model()
var = model.add_int_var(target=0)

# This is wrong!
model.add_dual_bound(max(var, 1))

# This is correct.
model.add_dual_bound(dp.max(var, 1))

The same thing applies to min().

Using If Statements with Condition

Values of expressions are not determined immediately: it is evaluated given a state inside the solver. Therefore, you cannot use if statements with Condition because the result of the if statement is determined immediately. Instead, you need to use if_then_else().

import didppy as dp

model = dp.Model()
var = model.add_int_var(target=0)

# This is wrong!
if var >= 1:
    model.add_dual_bound(1)
else:
    model.add_dual_bound(0)

# This is correct.
model.add_dual_bound((var >= 1).if_then_else(1, 0))

If you want to make a transition available only when a condition is satisfied, you should define it as a precondition.

import didppy as dp

model = dp.Model()
var = model.add_int_var(target=0)

# This is wrong!
if var <= 2:
    transition = dp.Transition(
        name='increment',
        cost=1 + dp.IntExpr.state_cost(),
        effects=[(var, var + 1)],
    )
    model.add_transition(transition)

# This is correct.
transition = dp.Transition(
    name='increment',
    cost=1 + dp.IntExpr.state_cost(),
    effects=[(var, var + 1)],
    preconditions=[var <= 2],
)
model.add_transition(transition)

Using Boolean Operators in Condition

When you want to take the negation, disjunction, and conjunction of Condition, you cannot use built-in boolean operators (not, or, and and) in Python. Instead, you need to use bitwise operators (~, |, and &).

import didppy as dp

model = dp.Model()
var = model.add_int_var(target=0)

# This is wrong!
model.add_base_case([(var >= 0 or not var >= 3) and var <= 2])

# This is correct.
model.add_base_case([(var >= 0 or ~(var >= 3)) & (var <= 2)])

Using a Table as a Nested List

A table in the model can be created from a nested list, but it is not a nested list. Use tuple as indices instead of nested indices.

import didppy as dp

model = dp.Model()
obj = model.add_object_type(number=2)
var = model.add_element_var(object_type=obj, target=0)

table = model.add_int_table([[1, 2], [3, 4]])

# This is wrong!
model.add_base_case([table[var][0] == 2])

# This is correct.
model.add_base_case([table[var, 0] == 2])

Using an Inappropriate Solver

The solvers provided by DIDPPy are not always applicable to all models. If a solver produces a wrong solution, it is likely that the model is not supported by the solver. Please refer to the solver selection guide and the API reference to check which solver supports which types of models.