Error: Parameters are currently only allowed to enter function such as exp, sin etc as exp(a), sin(b) etc.

[Muhammad Shahzad]

u = sdpvar(repmat(nu,1,N),repmat(1,1,N));

y_pred = (sdpvar(repmat(nx,1,N),repmat(1,1,N)))';

r = (sdpvar(repmat(ny,1,N),repmat(1,1,N)))';

constraints = [];

objective = 0;

for k = 1:N

 %x = A*x + B*u{k};

 y_pred{k} = simlssvm({input_train',output_train',type,6965.92075677792,0.0272260740857455,'RBF_kernel','orignal'},{alpha_pso,b_pso},u{k}');

 

 objective = objective + norm(y_pred{k}-(r{k}),1) + norm(R*u{k},1);

 constraints = [constraints, -1 <= u{k}<= 1 , -1 <=y_pred{k} <= 1];

end

%%         Now Implementation   %%

parameters_in = {[r{:}]};

solutions_out = {[u{:}], [y_pred{:}]};

controller = optimizer(constraints, objective,sdpsettings('solver','gurobi'),parameters_in,solutions_out);

%controller = optimizer(constraints, objective,[],[],[u{:}]);

[Muhammad Shahzad]

If i use "fmincon" instead of Gurobi then following error occurs:

Error using ldl

Out of memory. Type "help memory" for your options.

Error in formAndFactorKKTmatrix

Error in formAndFactorKKTmatrix

Error in computeTrialStep

Error in barrier

Error in fmincon (line 813)

    [X,FVAL,EXITFLAG,OUTPUT,LAMBDA,GRAD,HESSIAN] =

    barrier(funfcn,X,A,B,Aeq,Beq,l,u,confcn,options.HessFcn, ...

Error in callfmincon (line 91)

[xout,fmin,flag,output,lambda] =

fmincon('fmincon_fun_liftlayer',model.x0,model.A,model.b,model.Aeq,model.beq,model.lb,model.ub,callback_con,model.options.fmincon,model);

Error in optimizer/subsref (line 326)

                eval(['output = ' self.model.solver.call '(self.model);']);

[Johan Löfberg]

answered on the wiki 

(cannot use simlssvm with yalmip objects)

[Muhammad Shahzad]

Thank you, sir, I have changed "gurobi" to "fmincon" by using []. And reduced horizon from N=7 to N=2. now it's working but took a long time to run, even it's more than an hour, still, simulation is in progress.

N=2; what other option can i use please?

%%         Now Implementation   %%

parameters_in = {[r{:}]};

solutions_out = {[u{:}]};

controller = optimizer(constraints, objective,[],parameters_in,solutions_out);

implementedU = [];

for j =1 : 10

    

    y_reff = sin(y_red(j:j+1));

    U = controller{y_reff};

    implementedU = [implementedU;U(1)];

end

[Johan Löfberg]

it shouldn't work. If it appears to work, that's just a fact from the fact that optimizer assumes that the user defines a model which can be handled yalmip once the parameters are fixed, and if it doesn't the behavior is undefined

[Muhammad Shahzad]

thank you so much I understand now. Please suggest that is there any other alternative to apply MPC on problems like this?

[Johan Löfberg]

You baically have to solve black-box optimization problems (or sit down and derive the derivatives etc of your predictions), and then solve your favorite nonlinear solver. In other words, there is no good convenient way

[Muhammad Shahzad]

thankyou so much for your guidance sir.