Writing Optimization Code in Matlab - Resource Guide

[physicsCU]

OK, here is the situation.

I am supposed to write optimization code in Matlab to determine which of two missions an airplane should perform. There are three total, but one of them has been decided on. So I need to determine which of the other two I should do.

The problem is that I have never written this type of code before, so I have no clue what to do.

Could someone give me a good resource I can use as a starting point to writing this code?

 

[neurocomp2003]

um i think you need to put parameters/variables/criterion that your using so that people can be more descriptive. ie is it a airplane path like TSP/Hcycles problem?

 

[physicsCU]

ok, I need to minimize the weight as much as i can.

one mission involved flying up to 10 liters of water, and is scored by number of laps completed ^2, ie 5 laps = 25 for score

the other is carry 96 tennis balls for 2 minutes, and this can be any number, as long as all 96 fly. scored by 1/RAC where RAC = airframe + battery weight.

The plane already will carry 48 tennis balls and 4 liters of water.

 

[Antiphon]

There is an optimization function built in.

What you want to do is define a goal function. The MATLAB function will
minimize that goal function (If you want to maximize it, put a minus sign in front
of it.)

Then there are some number of unknowns. You goal function is a function of the
unknowns and MATLAB will find particular values for the unknowns so that the
goal function is a minimum.


Try "help fminsearch"

 

[physicsCU]

thanks!

is fminsearch in standard MATLAB or is it in a toolbox?

 

[PerennialII]

... if you have the optimization toolbox then you've a number of great tools available. Fminsearch is for unconstrained nonlinear optimization and is a standard function.

 

[physicsCU]

ok thanks.

sounds like i may want to consider using the optimization toolbox, especially because i need this code done in two weeks.

 

[PerennialII]

... and depending on the degree of nonlinearity and in case you've 'complex' constraints you may have to implement, the optimization toolbox rules.