Translated Title
Research on genetic crossover for VLSI standard cell placement by comparing
Translated Abstract
The success of genetic algorithm is its evolutionary mechanism, including crossover and muta- tion. Crossover has a decisive influence on overall performance of genetic algorithm, so it has become a key factor in the design of genetic algorithm for large-scale problem. We first briefly introduce VLSI stand- ard cell placement problem and its chromosome coding, present basic idea of four major kinds of crossover and their algorithm steps respectively, and improve the cycle crossover. By using standard test examples, we conduct experiments on the performance of these four kinds of crossover for further comparing, analyze the correlation of their characteristics and performance, and then summarize the idea for designing high- performance crossover. Experimental results of the improved method named length-limited cycle crossover verify the validity of this idea.
Translated Keyword
comparing
crossover
genetic algorithm
VLSI standard cell placement
Access Number
WF:perioarticalCASS_47990934
