JijModeling 2.1.0 Release Notes

JijModeling 2.1.0 Release Notes#

Feature Enhancements#

Random Instance Data Generation#

JijModeling 2 now supports random instance data generation via Problem.generate_random_dataset and Problem.generate_random_instance! Please refer to the API documentation for more details.

Example usage:

import jijmodeling as jm
import builtins
problem = jm.Problem("problem")
N = problem.Natural("N")
c = problem.Float("c", shape=(N,))
x = problem.BinaryVar("x", shape=(N,))
problem += jm.sum(N, lambda i: c[i] * x[i])
inputs = problem.generate_random_dataset(
    options={
        'N': {"value": builtins.range(10, 20)},
        'c': {"value": jm.generation.value.closed(-1.0, 1.0)}
         # You can also specify "size" for the range of jagged array dimension size.
    },
    seed=123 # omittable
)
assert set(inputs.keys()) == {"N", "c"}
inputs

{'N': 11,
 'c': array([0.9391445850004434, -0.06511935276658232, -0.7460324048682914,
        -0.3244370577929385, 0.9998145067841699, -0.24407534521565966,
        0.31329468992315457, 0.5220645290926087, -0.12919360384851108,
        0.30443086831711375, 0.531258381213684], dtype=object)}

Support for the `jm.Expression()` constructor#

The Expression class now provides a constructor, allowing you to explicitly convert any value that is convertible to Expression (i.e., ExpressionLike) into an Expression object.

Improved Type Hints#

The bundled type hints are now more accurate, making it easier for IDEs to provide assistance.

Bugfixes#

Bugfix 1: Division containing decision variables only on left-hand side now compiles#

In JijModeling 2.0.0, divisions where no decision variable appears on the right-hand side, e.g. x / 2, incorrectly triggered a compile-time error. With this release, division now compiles correctly as long as the right-hand side does not contain decision variables.

Bugfix 2: Index/value comparison inside `map` and `filter` now compiles#

There was a bug where code comparing indices and values inside nested map or filter would fail to compile. For example, the previous version produced an error like:

@jm.Problem.define("TestProblem")
def problem(problem: jm.DecoratedProblem):
    V = problem.Natural(ndim=1)
    W = problem.Natural()
    x = problem.BinaryVar( shape=(W,))
    problem += problem.Constraint(
        "constr",
        [jm.sum(x[j] for j in W if j <= i) == 1 for i in V],
    )
# TypeError: Traceback (most recent last):
# ...
#     9  |          [jm.sum(x[j] for j in W if j <= i) == 1 for i in V],
#                                              ^^^^^^

# Type Error: Instance for comparison operator not found for type natural and ElementOf[set(V)]

Since this version, this compiles correctly.

Bugfix 3: Indexing with tuples now works correctly#

In previous releases, there was a bug where indexing with tuples could cause the compiler to crash with a PanicException under certain conditions. With this release, the bug has been fixed, and indexing dictionaries with single tuples now evaluates correctly.

Bugfix 4: `Problem.infer()` now correctly converts to expressions#

In previous releases, Problem.infer() raised a runtime error if its argument was not a Expression object. With this release, arguments that can be converted to expressions (such as numeric values or Placeholder objects) are now converted to expressions before type inference.

Bugfix 5: Missing parentheses in formula output#

Previously, exponentiation of summation symbols and addition/subtraction within summation symbols were not properly enclosed in parentheses, leading to ambiguity in the formulas. This issue has been fixed.

Other Changes#

Problem now provides Problem.name and Problem.sense properties.