Re: How to find the minimum sum value: min(max sum value)

[Johan Löfberg]

Confusing as you seem to mix the terms max (take the max of a vector) and max (maximize a scalar function). Use max/min/minimize/maximize when you describe things

[bin sun]

Dear Johan,

I think my final objective function is to take the min of a vector. But in this vector, each element is a  max (maximize a scalar function).

So for arrangement one, I want to maximize a scalar function: sum1=I1+I2+I3. I should get 3*476 for here.

     for arrangement two, I want to maximize a scalar function: sum1=I1+I2+I3. I should get 3*495 for here.

    And so on...

Then I have a vector=[sum1;sum2;...;sum10]. Then I want to find the min value of this vector.

So in summary, my objective is to find the minimum element from 10 max(maximize a scalar function).

At first I want to use f=sum(I), but it's wrong, since it can't maximize a scalar function.

Then I want to use f=-sum(I), but save all results and use min(all results) to find the min of the vector. But I don't know how to save all results.

Or I may use sort and sumk.

Thank you very much for your time,

Best,

Bin

[Johan Löfberg]

I have no idea what you are trying to do or say.

If you want to minimize a scalar f, you use it as objective. If you want to maximize it, you use f.

If you want to minimize max(q), you use max(q) as objective. If you want to maximize min(q), you use -min(q) as objective, etc.