how to define constraint on objective function in ga toolbox?

hi, this my optimization code:

lb=[0.01,0.0075,0.01];

ub=[0.2,0.07,0.07];

func=@efficiency;

[x,fvar]=gamultiobj(func,3,[],[],[],[],lb,ub);

plot(fvar(:,1),fvar(:,2),'o');

I have 3 decision variable sand two objective functions which are electrical and thermal efficiencies, so these efficiencies can not be higher that 100%, but ga calculated efficiencies higher than 100%. How can I define constraint (efficiencies <100%) on objective functions?

[Torsten]

Are lb and ub bounds on the efficiencies ? But you want to optimize the process - so why do you set upper bounds ?

And what value becomes greater than 1 ?

[Hannah Mohebi]

lb and ub are on my decision variables. I want to find maximum of efficiencies but the efficiency can not be more than 100%.

[Torsten]

And how do you get efficiency from your decision variables ? In other words: How does your objective function looks like ?

[Hannah Mohebi]

I wrote a function for my objective function, it is a physical proble, it is not a simple code.

[Torsten]

Can the numerical values for the efficiencies you calculate in the objective function become greater than 1 ? If yes: return min(1,efficiency) to ga or use nonlcon to limit efficiencies to 1 by the nonlinear constraint c(d1,d2,d3) <= 1.

[Hannah Mohebi]

I can not understand your meaning. Would you pleaseexplaine more? Would you please tell me how to write your answer in my code?

[Torsten]

Please answer the questions:

Does your objective function return efficiency as value to be maximized ?

o Yes o No

If yes, is it possible that the numerical value computed in the objective function for the efficiencies exceeds 1 ?

o Yes o No

If yes, define c(x) in function nonlcon such that efficiencies = c(x) <= 1.

[Hannah Mohebi]

The answer to first question: yes. Efficiency should be maximum, but considering that GA gets the minimum, I set the objective function equal to the negative efficiency.

The answer to second question: yes.

As I said, my objective function is not a simple function that I can define like c(x).

[Torsten]

Just copy what you calculate in the objective function to the function nonlcon to define c(x). What's the problem ?

[Hannah Mohebi]

No problem Thanks.

[Torsten]

To avoid duplicate code: Write a function where you compute the values of the objective function(s) and call this function from your objective as well as from nonlcon.