I have to use discrete model obtained by Koopman operator in my research, but I have only found various examples for continuous time systems. Could you please point me to Python code for MPC examples using discrete time systems?
Hello,
I just wanted to add that, in case you’re formulating a linear MPC with the Koopman operator-based model, this example might be useful as well.
It’s in MATLAB, but the commands should be very similar in Python.
Kind regards,
Josip
Thanks! It’s useful. Besides that, I have another question. If a continuous-time model is input, will Acados convert the continuous-time system into a discrete-time system through a discretization method (such as RK4)?
Thank you! It’s very useful. For a dynamic system, does the time interval ts when using a discrete-time model need to be equal to the simulation time tf / N_nodes when using a continuous-time model?
acados deploys an integrator corresponding to the options integrator_type, and sim_method_*.
Nonuniform time grids are supported by the acados integrators.
If you want to use discrete dynamics with nonuniform grids, you need to introduced a parameter for the time steps.
Have I understood you correctly that the sample time of the discrete model cannot be set by acados like it does for the RK integrators and must be either fed into the model at code generation time or updated as a runtime parameter by the user?
Will the integrator option “num_steps” have an effect on discrete “integrator” type as well or is it necessary to create more shooting nodes to get a finer sample time resolution of my discrete model?