改进混合遗传算法在建筑结构优化设计中的应用

针对遗传算法在迭代过程中经常出现未成熟收敛、振荡、随机性太大和迭代过程缓慢等缺点，提出引入转基因算子与单亲遗传算子，同时提出一种离散变量结构优化设计的三等分割算法，通过与遗传算法相结合并运用到初始群体形成和进化过程中，使两种算法既可相互独立地运算，又可彼此相互协调、共同作用．根据工程实际，充分考虑规范规定的约束条件和各项技术标准要求，建立离散变量结构优化模型．各种算法的优化结果对比表明，改进混合遗传算法具有省时、高效、局部搜索能力强和全局性好的特点。

Application of Improved Hybrid Genetic Algorithm to Optimized Design of Architecture Structures

In the iterative process of the standard genetic algorithm (SGA), there often appear premature convergence, oscillation, over-randomization and low iterative speed. To solve these problems, some improved measures including transgenic operator and one-parent genetic operator are proposed, and a three-equal-partition algorithm (TEPA) for the structural optimization with discrete variables is provided. This algorithm is then combined with genetic algorithm (GA) in the process of primal colony forming and colony evolving. Thus, the two algorithms independently operate, mutually harmonize and jointly play their roles. Moreover, on the basis of practical structure designs in engineering, a model of structural optimization with discrete variables is established by sufficiently consi-dering the constraint conditions stipulated by the norm and the demands for various engineering standards. The optimized results obtained by different algorithms show that the improved hybrid genetic algorithm is of the advantages of saving time, high efficiency, good local searching ability and excellent global convergence property.