A new genetic algorithm for continuous structural optimization
Document
Description
In this thesis, the author described a new genetic algorithm based on the idea: the better design could be found at the neighbor of the current best design. The details of the new genetic algorithm are described, including the rebuilding process from Micro-genetic algorithm and the different crossover and mutation formation.
Some popular examples, including two variable function optimization and simple truss models are used to test this algorithm. In these study, the new genetic algorithm is proved able to find the optimized results like other algorithms.
Besides, the author also tried to build one more complex truss model. After tests, the new genetic algorithm can produce a good and reasonable optimized result. Form the results, the rebuilding, crossover and mutation can the jobs as designed.
At last, the author also discussed two possible points to improve this new genetic algorithm: the population size and the algorithm flexibility. The simple result of 2D finite element optimization showed that the effectiveness could be better, with the improvement of these two points.
Some popular examples, including two variable function optimization and simple truss models are used to test this algorithm. In these study, the new genetic algorithm is proved able to find the optimized results like other algorithms.
Besides, the author also tried to build one more complex truss model. After tests, the new genetic algorithm can produce a good and reasonable optimized result. Form the results, the rebuilding, crossover and mutation can the jobs as designed.
At last, the author also discussed two possible points to improve this new genetic algorithm: the population size and the algorithm flexibility. The simple result of 2D finite element optimization showed that the effectiveness could be better, with the improvement of these two points.