241_pdfsam_VLSI TEST PRINCIPLES & ARCHITECTURES

241_pdfsam_VLSI TEST PRINCIPLES & ARCHITECTURES -...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
210 VLSI Test Principles and Architectures Algorithm 13 Simple_GA_ATPG 1: test set T =∅ ; 2: while there is improvement do 3: initialize a random GA currentPopulation; 4: compute ftness oF currentPopulation; 5: for i = 1 to maxGenerations do 6: add the best individual to test set T ; 7: nextPopulation =∅ ; 8: for j = 1 to populationSize/2 do 9: select parent 1 and parent 2 From currentPopulation; 10: crossover ±parent 1 ²parent 2 ²child 1 ²child 2 ³ ; 11: mutate ±child 1 ³ ; 12: mutate ±child 2 ³ ; 13: place child 1 and child 2 to nextPopulation; 14: end for 15: compute ftness oF nextPopulation; 16: currentPopulation = nextPopulation; 17: end for 18: end while 3 4 5 6 2 n 1 ± FIGURE 4.39 Roulette wheel selection. candidate parent individual. Finally, when comparing the effectiveness of roulette wheel with binary tournament selections, the notion of
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: selection pressure is nec-essary. Selection pressure is the driving force that determines the convergence rate of the GA population, in which the population converges to n identical (or very similar) individuals. Note that fast convergence may not necessarily lead to a better solution. Roulette wheel selection with replacement results in a higher selection pressure than binary tournament selection when there are some highly fit individ-uals in the population. On the other hand, when individuals’ fitnesses have a small variance, binary selection will apply a higher selection pressure....
View Full Document

Ask a homework question - tutors are online