PROCEEDINGS ARTICLE | February 10, 2006
Proc. SPIE. 6046, Fifth Symposium Optics in Industry
KEYWORDS: Fringe analysis, Genetic algorithms, Computer simulations, Neural networks, Genetics, Chemical elements, Algorithm development, Optimization (mathematics), Binary data, Fuzzy systems
In the last years, Soft computing techniques, such as Genetic Algorithms, Neural Networks and Fuzzy systems, have been applied in different science areas. In this work, two applications of Genetic Algorithms in engineering and optics are presented. The Genetic Algorithms are optimization, search and learning machine techniques, which work in a random way. To achieve the problem solution by using of Genetic Algorithms, an iterative process should be developed. First, the problem to solve is modelled in a mathematical way by establishing of a fitness or objective function. After, a random initial population of strings (chromosomes) codifying problem solutions is generated, which samples the search solution space of the fitness function. Then, offspring populations are generated from previous one by using genetic operators: selection, crossover and mutation. In the selection process, possible solutions are chosen depending on their fitness function value. Then, in the crossover procedure, string segments of pairs of solutions are exchanged to generate the next population. Finally, some parameters in the offspring population are changed by mutation with a low probability. Results of the application of Genetic Algorithms to solve fringe analysis and nesting in finite materials problems are presented.
