Using Simulink Fast Restart in a parallel computed optimization problem

Hi, I have a Simulink Model which works perfect when i turn on fast restart. I can change the Values of my parameters A1, A2 and i by hand and it does not need to recompile.

Now i want to use this model in a parameter optimization problem using parallel computing AND fast restart. It would save a lot of time to cut out the recompiling time. The Problem is parallel computing works great but the model compiles for every iteration.

I set up the optzimization like this

% Simulator set up with experiment data, multiple experiments
         Simulator = createSimulator(Exp(1));
         Simulator = fastRestart(Simulator,'On'); %Fast Restart ON??

% estimation function handle, sdo requires function with just one input
% argument
         estFcn = @(param) ObjectiveFun(param,Simulator,Exp);

% optimization options
            	
        parpool

        opt = sdo.OptimizeOptions;
        opt.Method = 'surrogateopt';
        opt.UseParallel = true;
        opt.OptimizedModel = 'Model';
        opt.ParallelFileDependencies = files;
        opt.MethodOptions.PlotFcn = 'surrogateoptplot';
        opt.MethodOptions.MaxFunctionEvaluations = Iterations;
        
% Optimization
        vOpt = sdo.optimize(estFcn,param,opt); 

I just tried running the same code without parallel computing (opt.UseParallel = false) and fast restart works.. but not with parallel computing

Answer 1

[S.L.]

Just if someone faces the same problem as me.

The problem was that my Simulink model was inside a Matlab Project. A part of the startup function of this project was to open the simulink model.

open('Test.slx')

When starting the optimization the startup function is called, which opens the model without fast restart, which the optimization function does not override..

I got arround adding those lines to the startup function:

open('Test.slx')
sim_obj = sdo.SimulationTest('Test');
sim_obj = fastRestart(sim_obj,'On');