CS 478 - Evolutionary Algorithms 1 - Neural Networks and Machine ...
CS 478 - Evolutionary Algorithms 1 - Neural Networks and Machine ...
CS 478 - Evolutionary Algorithms 1 - Neural Networks and Machine ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
There are other methods which lead to more diversity<br />
Rank selection<br />
– Rank order all c<strong>and</strong>idates<br />
– Do r<strong>and</strong>om selection weighted towards highest rank<br />
– Keeps actual fitness value from dominating<br />
Tournament selection<br />
– R<strong>and</strong>omly select two c<strong>and</strong>idates<br />
– The one with highest fitness is chosen with probability p, else the lesser is<br />
chosen<br />
– p is a user defined parameter, .5 < p < 1<br />
– Even more diversity<br />
Fitness scaling - Scale down fitness values during early generations.<br />
Scale back up with time. Equivalently could scale selection<br />
probability function over time.<br />
<strong>CS</strong> <strong>478</strong> - <strong>Evolutionary</strong> <strong>Algorithms</strong> 14