Print

Print


The paper below outlines the notion of "competitive fitness sharing".  
In short, the idea is that, in coevolution, individuals should be  
rewarded (given high fitness) both for being able to beat many  
opponents from the other population and for beating particularly hard- 
to-beat opponents. This is done by allocating a fixed amount of  
fitness for beating each opponent and dividing it across the  
individuals who do so.

http://www-cse.ucsd.edu/users/crosin/icgawww.ps

Does ECJ support this, or is it something I would have to write?

Peter Drake
http://www.lclark.edu/~drake/