沙河渡槽槽身结构优化设计

以遗传算法为全局搜索方法,以混合离散变量的直接搜索与查点法为局部搜索方法,对遗传算法改进而形成了混合遗传算法。研制开发了基于混合遗传算法的渡槽结构优化设计程序。对于沙河渡槽进行了优化设计,得到了U形断面4槽、2槽和矩形断面4槽、2槽共4种槽型方案。从中确定U形断面4槽方案为最优方案,并被设计单位采用。优化结果表明：U形渡槽的经济性和安全性优越于矩形渡槽;遗传算法是一个很有推广应用价值的优化设计方法。

Optimal Structure Design of the Shahe Aqueduct

The Genetic Algorithms (GA) is adopted as the global searching methods,and the direct search method of mixed discrete variables as the local searching method. Then the GAs is improved and the hybrid GA-based optimal design programs are developed,and used to optimize the Shahe aqueduct. Four aqueduct design plans are obtained: 2-line double U-sectioned, 2-line U-sectioned, 2-line double rectangle-sectioned, 2-line rectangle-sectioned. Finally the 2-line double U-sectioned aqueduct is chosen as the best design. Optimization results show that U-sectioned aque-duct is superior to rectangle-secrioned aqueduct in economic and safety aspects, and GAs are effective optimal design methods worth of application and dissemination.