The identification of appropriate reaction models is very helpful for developing chemical vapor deposition (CVD) processes. In this study, we have developed an automatic system to model reaction mechanisms in the CVD processes by analyzing the experimental results, which are cross-sectional
shapes of the deposited films on substrates with micrometer-or nanometer-sized trenches. We designed the inference engine to model the reaction mechanism in the system by the use of real-coded genetic algorithms (RCGAs). We studied the dependence of the system performance on two methods using
simple genetic algorithms (SGAs) and the RCGAs; the one involves the conventional GA operators and the other involves the blend crossover operator (BLX-alpha). Although we demonstrated that the systems using both the methods could successfully model the reaction mechanisms, the RCGAs showed
the better performance with respect to the accuracy and the calculation cost for identifying the models.
Journal for Nanoscience and Nanotechnology (JNN) is an international and multidisciplinary peer-reviewed journal with a wide-ranging coverage, consolidating research activities in all areas of nanoscience and nanotechnology into a single and unique reference source. JNN is the first cross-disciplinary journal to publish original full research articles, rapid communications of important new scientific and technological findings, timely state-of-the-art reviews with author's photo and short biography, and current research news encompassing the fundamental and applied research in all disciplines of science, engineering and medicine.