Abstract: In order to solve the problem of premature convergence or slow convergence of algorithms, which can be easily seen in the progress of composite laminates design and optimization, through combining multi population genetic algorithm and adaptive genetic algorithm, a new improved dual population genetic algorithm with a adaptive operator is proposed. Taking as an example of symmetric laminated for test, a Patran finite element initial model is established. And, with Matlab, the main program of genetic algorithm and data transfer procedures is compiled, by which the Nastran input and output files can be written and read. Then a fitness function based on Tsai-Wu failure criterion is established, and the laminate stacking sequence is optimized. Compared with the traditional genetic algorithm, the results show that the improved algorithm can improve the optimization efficiency obviously, and can effectively converge to the global optimal solution, consequently providing a new solution to the optimization problem of composite structures.

Key words: dual population genetic algorithm, composite laminate, stacking sequence