27.06.2013 Views

Evolution and Optimum Seeking

Evolution and Optimum Seeking

Evolution and Optimum Seeking

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Multidimensional Strategies 63<br />

(If necessary repeat this process until Gj(x (0 ) ) 0 for all j =1(1)m.)<br />

Set k =0.<br />

Step 1: (Re ection)<br />

Determine the index w (worst vertex) such that<br />

F (x (kw) ) = maxfF (x (k ) ) =1(1)Ng:<br />

Construct x = 1<br />

N;1<br />

NP<br />

=1<br />

6=w<br />

(k ) x<br />

<strong>and</strong> x 0 = x + (x ; x (kw) ) (over-re ection factor =1:3):<br />

Step 2: (Check for constraints)<br />

If m = 0, go to step 7 otherwise set j =1.<br />

If m 1 = 0 , go to step 5.<br />

Step 3: (Set vertex back into bounds for explicit constraints)<br />

Obtain g = Gj(x0 )=Gj(x0 `j).<br />

If g 0, go to step 4<br />

otherwise replace x0 `j x0 `j + g + " (backwards length " =10 ;6 ).<br />

If Gj(x0 ) < 0, replace x0 `j x0`j ; 2(g + ").<br />

Step 4: (Explicit constraints loop)<br />

Increase j j +1.<br />

If<br />

8<br />

><<br />

>:<br />

j m 1 go to step 3<br />

m 1 m go to step 7:<br />

Step 5: (Check implicit constraints)<br />

If Gj(x 0 ) 0, go to step 6<br />

otherwise go to step 8, unless the same constraint caused a failure ve times<br />

in a row without its function value Gj(x 0 ) being changed. In this case go to<br />

step 9.<br />

Step 6: (Implicit constraints loop)<br />

If j < m, increase j j +1 <strong>and</strong> go to step 5.<br />

Step 7: (Check for improvement)<br />

If F (x0 )

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

Saved successfully!

Ooh no, something went wrong!