基于混合遗传算法的随机结构可靠性优化设计

遗传算法（GA）是一种具有随机搜索技术的进化算法，但在运用过程中出现早熟、收敛速度慢、局部搜索能力差的缺点。对遗传算法中的遗传算子进行了改进，提出了遗传算法和最佳矢量法相结合的混合遗传算法，并引入了小生境技术。分析表明，基于小生境的混合遗传算法即发挥了最佳矢量法局部搜索能力强的特点，又结合了遗传算法全局搜索能力强的优点，使收敛性能大大改善，同时小生境技术的使用，避免了优化过程中局部最优解的出现，提高收敛速度。具体算例表明该混合遗传算法是一种高效的结构优化方法。

Reliability Optimization Design of Stochastic Structure Based on Hybrid Genetic Algorithm

Based on randomness appeared in stiffness, external load and strength of a structure system in applying, an optimized design method of a structure system is proposed by considering random factors. In the method, sensitivity analysis of the system reliability index is first carried out. Then, the reliability index is considered as a constraint function to establish an optimized model and structure weight is optimized. Here, hybrid genetic algorithm of the traditional genetic algorithm combined with the optimal vector method is presented, and the Niche technology is introduced to perform structural optimized design. Analytical results indicate that the hybrid genetic algorithm based on Niche technology not only has the strong local searching capacity of the optimal vector method but also has the strong general searching capacity of the genetic algorithm, so convergence performance is greatly improved. Calculated example indicates that hybrid genetic algorithm based on the Niche technology is a high-effective structural optimized method.