Trouble with Solver Error and Infeasible Initial Solution

[hai02562]

Dear Elena,

I've been working with your Benders Decomposition implementation and I'm keen to apply it to various problems. Your code works quite well, but I'm encountering a problem that it can't seem to resolve.

Here are the parameters I'm using in the __init__ function:

self.Nx = 1
self.Ny = 3
self.m = 5
self.n = 2
self.c = np.array([-14]).reshape((self.Nx, 1))
self.f = np.array([0, -18, 0]).reshape((self.Ny, 1))
self.A = np.array([[0], [0], [-4], [34], [0]]).reshape((self.m, self.Nx))
self.B = np.array([[0, -11, -18], [0, -19, 0], [0, 39, 42], [-31, 0, 0], [0, 0, 0]]).reshape((self.m, self.Ny))
self.b = np.array([-175, 0, 0, 0, 0]).reshape((self.m, 1))
self.D = np.array([[0, 19, 0], [0, 0, 21]]).reshape((self.n, self.Ny))
self.d = np.array([-101, 0]).reshape((self.n, 1))
self.y_init = np.array([1, 0, 1], dtype=int).reshape((self.Ny, 1))  # Initial feasible guess

However, when running the code, I receive the following error:

cvxpy.error.SolverError: Solver for Master problem failed. Please try a feasible initial solution guess y_init!

The initial solution I provided is actually the exact solution, which I discovered by creating a Pyomo model with these parameters and solving it with MOSEK.
Additionally, I receive the following error message:

cvxpy.error.SolverError: Solver 'MOSEK' failed. Try another solver, or solve with verbose=True for more information.

This doesn't make sense to me as I've also used MOSEK for the Pyomo solution.
I'm wondering if you might have any insights on how to fix this issue? I would greatly appreciate any help you can provide. Thank you very much!

Best regards,