Genetic Algorithm and objective function value

Hello everyone,

I was doing an optimization using GA (Genetic Algorithm). After nearly one hour and half, i obtained the optimized parameters and the minimum value of the objective function.

To be sure from the solution, i put the optimized parameter in the same model and here is the surprise ! I obtained another value of the objective function !!!

Has anyone encountered this problem and knew why ?

As shows the figure below the value of the objective function is 0.114106.

However, when i put the same parameters in the main code, i obtained 0.1253 as shows the figure below.

What do you think ?

Answer 1

[Star Strider]

If the objective funciton is the same as the fitness function (and I have no idea what you are doing), the result should be the same, and equal to the best fitness value. If they are different (for example in a curve-fitting problem where the norm of the difference between the data and the objective function is to be minimised), the objective function value would be expected to be different from the fitness function value.

[nado]

@Star Strider,

i totally agree with you. That is why i find it weird.

In my case, the objective function is the same as the fitness function. In fact, the model is written in the form of a script. I copy the script and write it in the form of a function to be able to use GA.

I think the problem is in the script and i should review it.

Thanks a lot !

[Star Strider]

As always, my pleasure!

I have ga display all the results at the end, including the final fitness value. I am not certain what you have it display, however that might help uncover whatever the discrepancy may be. Also, if you are using a hybrid function to ‘fine-tune’ the parameters at the end, that could explain the difference.

.