def test_lqr_optimal_policy_no_scipy(self, make_env): env = make_env() old_sp = lqr_solver.sp try: lqr_solver.sp = None # Force the solver to use the non-scipy code path. p, k, beta = lqr_solver.solve(env) finally: lqr_solver.sp = old_sp self.assertPolicyisOptimal(env, p, k, beta)
def test_lqr_optimal_policy(self, make_env): env = make_env() p, k, beta = lqr_solver.solve(env) self.assertPolicyisOptimal(env, p, k, beta)