基于遗传算法与牛顿法的联合算法在地表发射率反演中的应用

结合遗传算法全局高效搜索和牛顿法局部细致搜索的优势,充分利用一种算法的优点弥补另一种算法的不足,进而引入一种基于遗传算法和牛顿法的联合算法,并将联合算法应用于反演地表发射率的函数关系中.结果表明,联合算法中由遗传算法提供的初始值使得牛顿法下降的速度快,且很快趋于稳定,达到精度要求;而由任意初始值提供给牛顿法,目标函数下降到一定阶段后反而有所回升,然后才保持稳定,且经和联合算法迭代相同的次数后,目标函数的值仍然非常大,远远达不到要求.因此,从可行性、计算效率上看,联合算法均优于单纯的牛顿法,是一种性能稳定,计算高效的下降方法.

Joint Algorithm Based on Genetic Algorithm and Newton's Method in the Application of Land Surface Emissivity Retrieval

In this paper,Combined the global genetic algorithm(GA) to efficient global search and detailed search of the Newton's method local advantages,make full use of the advantages of an algorithm to make up for disadvantages of another algorithm,introduced a joint algorithm based on genetic algorithm and Newton's method.And the joint algorithm is applied to the surface emissivity function relationships retrieval.The results show that the initial value of the joint algorithm provided by the genetic algorithm makes fast decline of Newton's method,and was soon stabilized to achieve error requirement;and by an arbitrary initial value for Newton's method,the objective function has decreased to a certain stage after recovering somewhat and then remain stable times,and by the same number of joint algorithm iterations,the value of the objective function is still very large,far below standard.Therefore,from the point of view of feasibility and calculation efficiency,joint algorithm is superior to single Newton's method,It is a stable,computationally efficient descent method.