防洪优化调度多约束启发式逐步优化方法

A multi-constrained heuristic progressive optimality algorithm for flood control operation

决策者的经验和知识判断对实际洪水调度有着非常重要的影响,为此,将决策者预泄控制的实践经验和泄流状态持续性要求纳入模型约束条件中,提出了防洪优化调度多约束启发式逐步优化方法。该方法以最大削峰为控制目标,将水库泄流的一般操作原则概化为启发式信息,以洪水涨落和水位升降作为泄流增减的判别依据,与优化搜索相结合,采用启发式逐步优化算法求解。以水口水库为例,采用5个方案对所提方法进行了对比分析。结果表明,提出的方法可以充分考虑决策者的经验和知识判断,降低了预泄腾库水位过低造成的回蓄风险,避免了泄流过程的波动,得到符合实际洪水调度要求的计算结果。

The decision maker＇s experience and knowledge on judgments have a significant influence on the actual flood control operation. In this paper, the decision maker＇s experience on flood control pre-discharge dispatching and on the required persistent state of releasing flood waters are considered as constraints , and a multi-constrained heuris- tic; progressive optimality algorithm （MHPOA） for flood control operation is thus proposed. In the MHPOA, the gener- al operation rule for flood releasing is seen as heuristic information, which controls the operation of increasing or de- creasing discharge rates by judging the variation tendency of inflow and reservoir level. The heuristic information is combined with the progressive optimality algorithm to searching for the optimal operation of flood control with the ob- jective of maximal flood peak reduction. The proposed algorithm is tested for 5 different scenarios designed for optimi- zing the flood control operation in the Shuikou reservoir. The result shows that the MHPOA is able take a full account of decision maker＇s experience and knowledge on judgments, reduce the refill risk caused by the issue of too low reser- voir levels in the pre-discharge period, avoid the fluctuation of discharge, and obtain the satistactory solution.