基于混沌DE算法和PP多项式函数的洪灾评估  

Flood disaster evaluation based on chaotic differential evolution algorithm and projection pursuit polynomial function

作　　者：何耀耀[1] 周建中[1] 罗志猛[1] 卢有麟[1] 

机构地区：[1]华中科技大学水电与数字化工程学院,湖北武汉430074

出　　处：《人民长江》2010年第3期92-95,共4页Yangtze River

基　　金：国家重点基础研究发展计划973项目(2007CB714107);国家自然科学基金重点项目(50539140);科技部水利部公益性行业科研专项资助项目(200701008)

摘　　要：针对洪水灾情等级评估问题,建立了投影寻踪(PP)多项式函数模型,并提出一种基于分段线性混沌映射的混沌差分进化(CDE)算法优化投影指标函数,同时采用罚函数法处理约束条件,求出了最佳投影方向。通过灾情等级表各指标边界值的投影求出多项式函数的系数,获得一个评价洪水灾害等级的多项式函数,从而实现连续型洪灾等级的划分。2个实例计算的结果以及与其他现有的模型和算法的对比表明,该方案不仅有效,且更为简单实用,具有广泛的应用性。In the light of the problem of flood disaster grade evaluation,we establish the projection pursuit（pp） polynomial function model,and propose a chaotic differential evolution（CDE） algorithm based on piecewise linear chaotic mapping to optimize projection index function,and adopt penalty function method to deal with constraint condition for detecting optimal projection direction.The coefficients of polynomial function are solved by projecting the index boundary values in flood disaster grade table so as to obtain a polynomial function of flood grade division and realize continuous grade division of flood disaster.The calculation results of two examples and the comparison with other existing models and algorithms show that this scheme is effective,simple and has wide application.

关 键 词：混沌优化 CDE算法 投影寻踪 多项式函数 洪灾评估 

分 类 号：P426.616[天文地球—大气科学及气象学]