分片近似法求解数学物理方程反问题
Partitioned Approximating Method of Solving Indirect Problems of Mathematical Physics Equation
指出了基于优化理论的数理方程反演方法的局限性。分析了直接法的优点和缺陷,从弥补直接法的缺陷出发,借鉴有限元的思想并利用双差准则的优势,提出了数理方程反演问题的一种新的直接解法----分片拟合法。论述了该方法的基本原理和需要解决的关键问题――近似函数的收敛性和系数确定过程中的计算量以及相应的对策――采用分片拟合以降低计算量,采用双差准则以防止近似函数的发散。随后列举了一个常微分方程反演的例子以说明该方法的构造和求解过程。与基于优化理论的反演方法相比,本文提出的方法不需进行微分方程的正演过程(数值正演或解析正演),不需经过参数的试算过程,即可直接由观测数据一次求出所需反演的参数。数值实验的结果表明了本文所提算法的精确性和有效性。
he limitation of inversion method that based on optimization theory is pointed out. The advantage and disadvantage of direct method is analized. In oder to cover the shortage of the direct methods, a new method---partitioned approximating method of solving the indirect problems of mathematical physics equation is proposed after borrowing ideas from finite element method and making use of double difference criterion’s merits. The priciples of the method and the key problems (approximating function’s convergence and computation time of dertermining the coefficient) that need to solve is discussed. The corresponding countermeasure is presented: adopt partitioned approximation to avoid the computing dificulty and adopt double difference criterion to avoid the approximating function’s diverging. Then an example of ordinary differential equation’s inversion is listed to instruct the constructing process and solving process of the method. Compared to the inversin method that based on optimization theory, the proposed method of this paper does not need to solve the differential equation of the inverse problem and does not need the parameter’s pilot calculation and the parameter that need to inverse can be got by once-through calculation. The numerical experiment’s results indicate the proposed method’s precision and validity.
童海滨
数学物理学
数学物理方程 反演 优化 分片近似法 双差准则
mathematical physics equation inversion optimization approximating method double-difference criterion
童海滨.分片近似法求解数学物理方程反问题[EB/OL].(2006-04-10)[2025-08-03].http://www.paper.edu.cn/releasepaper/content/200604-123.点此复制
评论