Problems with defining upper and lower bounds

Hi!

I'm trying to run GA with

[x fval] = ga(@objectiveFunction, nvars, [],[],[],[],lower, upper,[], integers, options)

Where:

nvars = 2, 
lower = [3 3]
upper = [13 10]
integers = []

Problem is that the parameters sent to the objective function from the GA sometimes are outside the bounds i've specified.

For testing the objective function now only writes it's input to the console and returns a fixed number.

Here's a part of the console output:

param =

      0.7289    3.1917

param =

      0.9995    7.7736

param =

     -1.2105    3.4438

param =

      7.9082    6.2424

As you can see several of the values are outside the allowed range. What have i done wrong?

Best regards.

Daniel

Answer 1

[Daniel]

I solved it by using @mutationadaptfeasible as mutation function!

Answer 2

[Sean de Wolski]

The final output value is inside the range though, right? I believe ga() is showing every attempt, even if the attempt falls outside of the constraints, but it will converge to something inside the constraints.

[Daniel]

Actually, even the end result is outside the bounds unfortunately.

[Sean de Wolski]

What options are you using?

param above, looked like it was working its way in.

[Daniel]

options = gaoptimset(@ga);

options.Vectorized = 'off';

Basically default options.