fix(simplex): raise ValueError instead of returning empty dict when max iterations reached

Fixes #14584

The run_simplex() method now raises a descriptive ValueError when
the maximum iteration limit (Tableau.maxiter) is reached, instead of
silently returning an empty dict. This allows callers to distinguish
between a converging solution and a cycling/unbounded problem.

Author: MiniMax Agent

linear_programming/simplex.py

@@ -302,7 +302,13 @@ def run_simplex(self) -> dict[Any, Any]:
                 self.tableau = self.change_stage()
             else:
                 self.tableau = self.pivot(row_idx, col_idx)
-        return {}
+
+        max_iterations = Tableau.maxiter
+        raise ValueError(
+            "Simplex did not converge within "
+            + str(max_iterations)
+            + " iterations. The problem may be cycling or unbounded."
+        )
 
     def interpret_tableau(self) -> dict[str, float]:
         """Given the final tableau, add the corresponding values of the basic