Failure in initial objective function evaluation. FMINCON cannot continue in s function

I am using a level 2 s function to perform fmincon and I need to limit the rate of change of output. I have stored the output using Dwork vectors. I want to create a constraint where the difference between the current output value and previous output value should be less than 20.

function ECMSsf2(block)

setup(block);

%endfunction

function setup(block)

%% Register number of input and output ports

block.NumInputPorts = 2;

block.NumOutputPorts = 3;

%% Setup functional port properties to dynamically

%% inherited.

block.SetPreCompInpPortInfoToDynamic;

block.SetPreCompOutPortInfoToDynamic;

block.InputPort(1).DirectFeedthrough = false;

%% Set block sample time to inherited

block.SampleTimes = [-1 0];

%% Set the block simStateCompliance to default (i.e., same as a built-in block)

block.SimStateCompliance = 'DefaultSimState';

%% Register methods

block.RegBlockMethod('PostPropagationSetup',@DoPostPropSetup);

block.RegBlockMethod('InitializeConditions',@InitConditions);

block.RegBlockMethod('SetInputPortSamplingMode', @SetInpPortFrameData);

block.RegBlockMethod('Outputs', @Output);

block.RegBlockMethod('Update', @Update);

function DoPostPropSetup(block)

%% Setup Dwork

block.NumDworks = 1;

block.Dwork(1).Name = 'x0';

block.Dwork(1).Dimensions = 1;

block.Dwork(1).DatatypeID = 0;

block.Dwork(1).Complexity = 'Real';

block.Dwork(1).UsedAsDiscState = true;

end

function InitConditions(block)

% Initialize Dwork

block.Dwork(1).Data = 0;

end

function SetInpPortFrameData(block, idx, fd)

block.InputPort(idx).SamplingMode = fd;

block.InputPort(idx).SamplingMode = fd;

block.OutputPort(1).SamplingMode = fd;

block.OutputPort(2).SamplingMode = fd;

block.OutputPort(3).SamplingMode = fd;

end

function Output(block)

%constants initialization

Pbatt_char=5000;

SOC_min=0.2;

SOC_max=0.90;

Pfc_min=1000;

Pfc_max=10000;

Pbatt_max=17000;

%define Matrix Aeq

Aeq=[0 1 0;1 0 1];

%define Matrix beq

mu = 0.6;

beq=[(1-2*mu*((block.InputPort(2).Data-0.5*(SOC_max+SOC_min))/(SOC_max+SOC_min))); block.InputPort(1).Data];

%define boundary conditions

lb=[Pfc_min, 0, -Pbatt_char];

ub=[Pfc_max, 100, Pbatt_max];

%defining initial conditions

x0 = [0 0 0];

options = optimoptions('fmincon','Algorithm','active-set','Display','off','MaxFunctionEvaluations',1000,'MaxIterations',100);

[y,fval] = fmincon('OF_ECMS',x0,[],[],Aeq,beq,lb,ub,'Nonlinearequations',options); %#ok<*ASGLU>

Pfc=y(1); Pbatt=y(3); alpha=y(2);

block.OutPort(1).Data = Pfc;

block.OutPort(2).Data = Pbatt;

block.OutPort(3).Data = alpha;

end

function Update(block)

block.Dwork(1).Data = block.Outport(1).Data;

end

end

end

and the objective function is

function f = OF_ECMS(y(1),y(2),y(3));

f = (y(1)+y(2)*y(3));

end

and the constraints are

function [c, ceq] = Nonlinearequations(y,PFCold)

c = [];

ceq = [];

PFCold = block.Dwork(1).Data;

% Ramp constraints which are inequality constraints

c(1) = PFCold - y(1) + 20;

c(2) = y(1) - PFCold - 20;

end

Answer 1

[Matt J]

Both your objective function and constraints are linear, so you should be using linprog. In fact, you nonlinear constraints are simple bounds on y(1). So, do instead,

  lb=[Pfc_min, 0, -Pbatt_char];
  ub=[Pfc_max, 100, Pbatt_max];
  
  PFCold = block.Dwork(1).Data;
  lb(1)=max(lb(1), PFCold-20);
  ub(1)=min(ub(1), PFCold+20);
  
  
  [y,fval] = linprog([1,1,1], [],[],Aeq,beq,lb,ub);

[Matt J]

As a side note, however, if you were to use your original fmincon attempt, it should look like this:

nonlcon=@(y) Nonlinearequations(y,block);
[y,fval] = fmincon('OF_ECMS',x0,[],[],Aeq,beq,lb,ub,nonlcon,options);


function f = OF_ECMS(y) %<----fix
 f = (y(1)+y(2)*y(3));
end

function [c, ceq] = Nonlinearequations(y,block)%<----fix
    c = [];
    ceq = [];
   PFCold = block.Dwork(1).Data;
    

    c(1) = PFCold - y(1) - 20;  %<----change to -20
    c(2) = y(1) - PFCold - 20;
end

[Walter Roberson]

I really recommend using

[y,fval] = fmincon(@OF_ECMS,x0,[],[],Aeq,beq,lb,ub,nonlcon,options);

instead of the quoted name 'OF_ECMS' . Using the quoted name is supported by fmincon, but it has the disadvantage that the function must exist on the MATLAB path -- so in the case of providing your own function, that there has to be a separate .m function by that name. When you use a quoted name, you cannot refer to a function defined inside the same script or function as your other code.

[aditya deepak]

Thank you for your reply. There is a requirement that I have to use fmincon. I have seperate m files for constraint and the objective function. I would like to define both functions in the same script. Are there any examples. Thank you in advance

[aditya deepak]

I am getting the following error

Undefined function 'OutPort' for input arguments of type 'Simulink.MSFcnRunTimeBlock'

[Matt J]

I would like to define both functions in the same script. Are there any examples.

See Local Functions