设计洪水频率曲线的粒子群优化适线法研究

应用现代智能优化算法，按照不同适线准则进行了设计洪水的频率计算。根据我国水利水电工程设计洪水计算规范，应用粒子群算法按照离（残）差平方和最小，离（残）差绝对值最小和相对离（残）差平方和最小准则，进行了设计洪水频率曲线参数的优化估计研究。通过实例验证，认为粒子群算法进行设计洪水频率曲线参数估计优于传统的经验适线法。与传统优化方法相比，粒子群算法对优化目标函数要求低，避免了大量的求导计算，是一种推求设计洪水频率曲线统计参数的新途径。

Curve Fitting Method of Design Flood Frequency Based on Particle Swarm Optimization

The purpose of the study was to compute design flood frequency by the optimized algorithm. According to the regulations of design flood frequency computation for water resources and hydropower projects, a curve fitting method of design flood frequency was presented by using the particle swarm optimization and applying the WLS, ABS, and WLS. The method was then applied to the practical computation of design flood frequency. Results indicate that the method presented in the study is better than the empirical fitting method. Comparing with other conventional optimization methods, it requires a lower target function and need not compute the differential coefficient. The method is a new way of inferring statistical parameters of design flood frequency curve.