Feature: MatrixVariable supports numpy broadcast feature (#1092)

* Add broadcast test for matrix constraints

* Feature: support matrix broadcast

* lint via ruff

* Update CHANGELOG.md

* Update CHANGELOG.md

* Update CHANGELOG.md

* Update CHANGELOG.md

* miss launch optimization

* Set up bound and target

* lint codes

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update CHANGELOG.md

---------

Co-authored-by: João Dionísio <57299939+Joao-Dionisio@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Author: 3 people

CHANGELOG.md

@@ -4,6 +4,7 @@
 ### Added
 - Added possibility of having variables in exponent.
 - Added basic type stubs to help with IDE autocompletion and type checking.
+- MatrixVariable comparisons (<=, >=, ==) now support numpy's broadcast feature.
 ### Fixed
 - Implemented all binary operations between MatrixExpr and GenExpr
 - Fixed the type of @ matrix operation result from MatrixVariable to MatrixExpr.

src/pyscipopt/matrix.pxi

@@ -28,15 +28,14 @@ def _matrixexpr_richcmp(self, other, op):
         else:
             raise NotImplementedError("Can only support constraints with '<=', '>=', or '=='.")
 
-    res = np.empty(self.shape, dtype=object)
     if _is_number(other) or isinstance(other, Expr):
+        res = np.empty(self.shape, dtype=object)
         res.flat = [_richcmp(i, other, op) for i in self.flat]
 
     elif isinstance(other, np.ndarray):
-        if self.shape != other.shape:
-            raise ValueError("Shapes do not match for comparison.")
-
-        res.flat = [_richcmp(i, j, op) for i, j in zip(self.flat, other.flat)]
+        out = np.broadcast(self, other)
+        res = np.empty(out.shape, dtype=object)
+        res.flat = [_richcmp(i, j, op) for i, j in out]
 
     else:
         raise TypeError(f"Unsupported type {type(other)}")

tests/test_matrix_variable.py

@@ -1,10 +1,4 @@
 import operator
-import pdb
-import pprint
-import pytest
-from pyscipopt import Model, Variable, log, exp, cos, sin, sqrt
-from pyscipopt import Expr, MatrixExpr, MatrixVariable, MatrixExprCons, MatrixConstraint, ExprCons
-from pyscipopt.scip import GenExpr
 from time import time
 
 import numpy as np
@@ -22,10 +16,10 @@
     cos,
     exp,
     log,
-    quicksum,
     sin,
     sqrt,
 )
+from pyscipopt.scip import GenExpr
 
 
 def test_catching_errors():
@@ -525,3 +519,16 @@ def test_matrix_matmul_return_type():
     y = m.addMatrixVar((2, 3))
     z = m.addMatrixVar((3, 4))
     assert type(y @ z) is MatrixExpr
+
+
+def test_broadcast():
+    # test #1065
+    m = Model()
+    x = m.addMatrixVar((2, 3), ub=10)
+
+    m.addMatrixCons(x == np.zeros((2, 1)))
+
+    m.setObjective(x.sum(), "maximize")
+    m.optimize()
+
+    assert (m.getVal(x) == np.zeros((2, 3))).all()