I am using the matlab interface for code generation. I got special case for which the output status is "0" while most of all output vectors are NaN. Is there is a way to "reset" the solver online or get rid of theses NaN. The solver cannot continue when I reach this case and keep the NaN values for the output.
Best regards,
Jonathan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I suggest you change another initialization. Or you may check if your problem is suitable for ACADO. I have encounter this situation once. It was because the gradient of the initial point is so big that ACADO will give NaN output.
May this help you
Yutao
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have the same question and i couldn't find any solution... i just need to initialize in the actual state the failed solver, exactly as it does at the start of simulation. I need a sort of "restart from the very beginning of the process". the state in which the solver would initialize is surely a feasible condition (if i run the model from that condition the solver doesn't fail).
i've looked for a solution for days for such a trivial action... is anybody abe to help me? i'm using the make_custom_solver_sfunction.m
Best Regards,
Marco
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I am using the matlab interface for code generation. I got special case for which the output status is "0" while most of all output vectors are NaN. Is there is a way to "reset" the solver online or get rid of theses NaN. The solver cannot continue when I reach this case and keep the NaN values for the output.
Best regards,
Jonathan
I suggest you change another initialization. Or you may check if your problem is suitable for ACADO. I have encounter this situation once. It was because the gradient of the initial point is so big that ACADO will give NaN output.
May this help you
Yutao
Hi,
I have the same question and i couldn't find any solution... i just need to initialize in the actual state the failed solver, exactly as it does at the start of simulation. I need a sort of "restart from the very beginning of the process". the state in which the solver would initialize is surely a feasible condition (if i run the model from that condition the solver doesn't fail).
i've looked for a solution for days for such a trivial action... is anybody abe to help me? i'm using the make_custom_solver_sfunction.m
Best Regards,
Marco