27.06.2013 Views

Evolution and Optimum Seeking

Evolution and Optimum Seeking

Evolution and Optimum Seeking

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

136 <strong>Evolution</strong> Strategies for Numerical Optimization<br />

Making use of the relation (see Ryshik <strong>and</strong> Gradstein, 1963)<br />

one nally obtains<br />

Zp<br />

y=0<br />

erf( y) dy = p erf( p)+ exp(; 2 p2 ) ; 1<br />

p<br />

~p(jx`ij b) = erf<br />

p !<br />

2 b<br />

+ 1 p<br />

2 b<br />

In the following we refer to this expression as item v.<br />

v =~p(jx`ij b)<br />

"<br />

exp<br />

; 2 b2<br />

2<br />

!<br />

#<br />

; 1<br />

(5.24)<br />

With the above de nition of v, the total probability that a descendant N` is feasible, i.e.,<br />

that it satis es all the constraints, is<br />

pf eas =<br />

<strong>and</strong> the probability that N` is lethal is<br />

nY<br />

i=2<br />

= v n;1<br />

~p(jx`ij b)<br />

pleth =1; pf eas =1; v n;1<br />

Only non-lethal mutants come into consideration as parents of the next generation. Hence,<br />

instead of w(s` = s0 )wemust insert into Equation (5.15) the expression<br />

w(s` = s 0 ) pf eas = 1<br />

p 2<br />

<strong>and</strong> instead of p(s`

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!