FTOCP function

[tuanpfiev]

Hi Urosolia,
I am running the program and get this error:

=======================================
Time step: 1, Iteration: 2
Error using * (line 49)
Multiplying NaN with an SDPVAR makes no sense.

Error in FTOCP (line 55)
Cost = Cost + Qfun*lambda;

Error in LMPC (line 17)
[xPred, uPred ] = FTOCP(x_LMPC(:,t), N, Q, R, Qfun, SS, A, B, X, U);

Error in main (line 32)
[ x_LMPC, u_LMPC, SS, Qfunction, IterationCost] = LMPC(x0, SS, Qfun, A, B, Q, R, ...

=======================================

Could you please help me resolve this issue?
Thank you in advance.
Tuan

[urosolia]

Have you tried to update your MPT3 (https://www.mpt3.org/Main/Installation)?

[tuanpfiev]

Thank you so much for your reply. The optimization was failed and it caused the issue. By the way, I am really interested in your research on Robust/Adaptive LMPC and the one on the racing car. I kindly ask if you could share the source codes of those studies.

[Hidenver2016]

Hi Urosolia,

I also have similar problem and I updated my MPT3 as your said above.

Time step: 2, Iteration: 1, Cost: 587.737135646650586
Error using == (line 15)
Error using - (line 52)
Adding NaN to an SDPVAR makes no sense.

Error in FTOCP (line 26)
Constraints = [x_t == x{1}];

Error in LMPC (line 50)
[~, uPred ] = FTOCP(x_LMPC(:,t), N, Q, R, Qfun, SS,...

Error in main (line 56)
[ x_LMPC, u_LMPC, x_cl, u_cl, IterationCost, SS] = LMPC(x0, x_cl, u_cl, IterationCost, A, B, Q, R, ...

Could you please help me on this?
Thank you very much!

Bests,
Qiang

[urosolia]

Can you check if the solvers are correctly installed? Are you able to solve an optimization problem with GUROBI and YALMIP?

[Hidenver2016]

Thank you very much for your reply. I change LMPC_options.solver = 'quadprog'; right now it works.

[urosolia]

I always recommend GUROBI for high precision and speed. For academic purposes the license is free!

[Hidenver2016]

Thank you very much!

[Ghui-Yang]

Thank you very much for your reply. I change LMPC_options.solver = 'quadprog'; right now it works.

The “gurobi” solver needs to be downloaded and installed in Matlab enviroment.