在防洪决策流程中,灾情预测与评估扮演着十分重要的角色,根据所处的不同阶段,可以划分为灾前预测、灾中决策和灾后评估.本文设计了一种灾情预测与评估的智能系统.将灰色关联度分析法、BP神经网络、基于免疫遗传算法(IGA)的RBF神经网络三种方法应用到灾情评估系统中.仿真结果表明基于IGA的RBF神经网络有较高的准确性,因而有较好的实用性.最后采用灰色关联度分析法和IGA—RBF网络模型,给出了洪水灾度评估的系统实现.
Prediction and evaluation of flood disaster plays an important role in flood control decisionmaking. The system can be divided into three modules of prediction before disaster, decisionmaking during disaster, and evaluation after disaster according to different stages. In this paper, an intelligent system for disaster prediction and evaluation is designed. Three intelligent algorithms,which are gray relational analysis, BP neural network, and IGA (immune genetic algorithm)-RBF (radius base function) neural network, are proposed and applied to disaster evaluation. Simulation results show that IGA-RBF has high precision and excellent performance, so it is feasible to use it in practical applications. Finally, we provide the implementation of disaster evaluation system with these algorithms.