基于遗传算法和并行组合模拟退火算法的洪水过程缩放模型研究

Study on flood zoom model based on genetic algorithm and parallel recombination simulated annealing algorithm

针对泄洪风险和施工导流风险计算对洪水缩放的要求和传统洪水同频率放大方法中手工修匀任意性较大等方面的不足,提出采用计算机优化方法来实现洪水过程的自动放大。在满足洪峰流量约束和分时段洪量约束条件下,本文建立了以洪水过程模式尽量相似为目标的洪水过程放大优化模型,并采用具有全局搜索能力的遗传算法和并行组合模拟退火算法求解该模型。通过实例计算可以看出,这两种算法计算结果均能较好的满足洪峰洪量约束要求,并有效保持了典型洪水的模式,避免了人工修匀的任意性。

Aiming at the flood zoom demand of flood discharging risk, construction diversion risk and the deficiency of the traditional same-frequency flood amplification method, a new optimal amplification method for the flood process realized by the computer is presented in this paper. The objective of the optimal model is to keep the flood process mode similar and subject to restrictions of the actual peak flow discharge and flood volume in different period of time, the GA and PRSAA that have global optimal capabilities are used to solve the model in this paper. From the case study, it can be seen that the results of two algorithms can satisfy the restrictions of the actual peak flood discharge and flood volume and keep the mode of the typical flood effectively and avoid the manual randomicity.