gaplotpareto not supported for this algorithm

Hi,

I am running gamultiobj for an optimization of a simulation. I ran it with:

options = optimoptions('gamultiobj','PlotFcn',@gaplotpareto);

rng default % For reproducibility
fun = @CostFunctionOptim_Pareto;
[x,fval,exitflag,output] = gamultiobj(fun,4,A,b,Aeq,beq,lb,ub,nonlcon,options);

I cannot share the cost function because it's really long and requires external software, but I do not expect it is the issue because I have used it for optimziation before with patternsearch and ga.

Details about the cost function:

• Four design variables
• 3 Objectives to be minimized
• Runs external software, gets data, computes a mass, compliance and penalty based on a barrier method equation
• Objectives are to minimize mass, compliance and penalty

I get a graph for the pareto front saying:

Would anyone be able to suggest why this is happening?

[鹏蕾 李]

your target function definition is wrong.

function [obj1,obj2]=objective(x) %wrong

function obj=objective(x)

obj=[obj1,obj2] %Right

Answer 1

[Gifari Zulkarnaen]

Have you tried this advice?

https://www.mathworks.com/matlabcentral/answers/63086-gaplotpareto-for-3-objective-functions

[Ayman Al-Sukhon]

Thank you, sorry for missing that.

Answer 2

[Modess Seyednezhad]

I have the same issue and that Link was not usefule for me. Anyone can help me?

Answer 3

[Modess Seyednezhad]

My function is:

function COP = simpleMultiObjective(x)

% both dT and I0 & COP are vectors

% x(1)=dT;

% x(2)=I0;

% based on manufacturer data sheet

Vmax = 16.1; % unite : volt

Imax = 8; % unite : A

dTmax = 71; % unit : Kelvin

Th_nom = 30+273.15;

Th = 30+273.15; % the hot side temperature of 30 C

Tc = Th-x(1); % obtained from the chart

S = Vmax/Th_nom;

K = (Th_nom-dTmax)*(Vmax*Imax)/(2*Th_nom*dTmax);

R = (Th_nom-dTmax)*Vmax/(Th_nom*Imax);

Qc = S.*x(2).*Tc - (x(2).^2 .* R )/2 - K.*x(1);

Qh = S.*x(2).*Th + (x(2).^2 .* R )/2 - K.*x(1);

Qte = abs(Qh-Qc);

COP = Qc./Qte;

end

and the main function is :

close all ; clear ;clc;

fitness = @simpleMultiObjective;

nvar = 2;

Lb = [1, 3];

Ub = [19.2, 6.27];

options = optimoptions('gamultiobj','Display','iter', 'MaxGenerations',200,...

'PopulationSize',10,'FunctionTolerance',1e-4 ,'PlotFcn',@gaplotpareto);

[xf,fval,exitflag] = gamultiobj(fitness,nvar,[],[],[],[],Lb,Ub,options);

I have the same issue and cannot see data on the ParetoPlot.

Any suggestion?