How to get the cost value of the current solution at each horizon step?

Remi · October 13, 2025, 1:45pm

Hi

I’m using acados python and try to implement lexicographic optimization MPC.
For this, I would need the cost result of the solution for each horizon step. But as far as I know, it’s only possible to get it for the whole solution (using get_cost). Or it is?

kaethe · October 14, 2025, 8:56am

Hi

You can try to use AcadosCostConstraintEvaluator. It’s a bit of a prototype implementation though.

Easiest is just to build the stage cost function again yourself.

Best, Katrin

Remi · October 14, 2025, 12:25pm

Hi,
Thank you for your answer.
Would you have an implementation example? I couldn’t find any documentation.

Or otherwise a way to access the function from the AcadosOcpSolver?

Best, Rémi

kaethe · October 14, 2025, 1:59pm

Here is an example using the evaluator

github.com/acados/acados

examples/acados_python/evaluation/minimal_example_evaluation.py

ec59717be


      
          # ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
          # POSSIBILITY OF SUCH DAMAGE.;
          #
          import sys
          
          sys.path.insert(0, '../getting_started')
          
          from matplotlib import pyplot as plt
          
          from acados_template import AcadosOcp, AcadosOcpSolver, AcadosSimSolver, AcadosSim, ACADOS_INFTY
          from acados_template.mpc_utils import AcadosCostConstraintEvaluator
          from pendulum_model import export_pendulum_ode_model
          from utils_eval import plot_pendulum_eval
          import numpy as np
          import scipy.linalg
          import math
          from casadi import vertcat, SX
          
          # Define constraints to make evaluation more challenging
          constraint_par = {'omega_dot_min_1': -4,
                            'omega_dot_min_2': -6,