Fitness Normalization

[Celso França]

Hello,

My fitness implementation, looks like:

fitness = surprise + efficiency

Efficiency is a value in an interval between 0 and 1. 

However, the surprise value can assume any value great than zero (actually very big values can occur). 

These two elements must to have the same weigh in equation and, to reach that, I have to normalize the surprise for current entire population before sum with efficiency value.

Do you have any idea to make this in the framework?

Thank you so much!

Live Long And Prosper

Celso França.

[Paul]

You could apply a nonlinear transformation to surprise. So, for instance, change fitness to f(surprise; lambda) + efficiency where f(x; lambda) = 1 - exp(-lambda * x). (I'm assuming here that higher surprise values should yield greater fitness.) Other functions are possible; you just want an increasing function with range [0, 1] (give or take the endpoints).

[Celso França]

Hello Paul,

Thank you so much for your answer. I've been very happy since I saw it, because this solution will work perfectly in my research.

Over again, thank you!

Live Long and Prosper.

Celso França.