项目名称: 非局部模型的自适应算法研究
项目编号: No.11201462
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 数理科学和化学
项目作者: 赵旭鹰
作者单位: 中国科学院数学与系统科学研究院
项目金额: 22万元
中文摘要: 非局部特性在真实世界中是普遍存在的。因此,很有必要研究相应的非局部模型,包括非局部材料模型和非局部扩散模型,这将具有重大的实际意义。如何快速有效的求解非局部模型是当今科学计算领域一个重要的前沿课题。自适应有限元方法作为求解偏微分方程最有效的数值方法之一,是当前科学计算研究的热点。因此,本课题的主要目的是研究非局部模型的自适应有限元算法和相关数值分析,这对于探索非局部模型的高效数值求解算法,以及开发非局部模型的数值计算软件包,都有重要的指导意义。本项目主要内容包括:研究含奇异积分核的非局部模型的有限元后验误差分析及自适应算法的收敛性分析;提出"投影型"后验误差分析方法并研究含超强奇异积分核的非局部模型的自适应有限元方法;建立"投影型"后验误差分析方法的一般性框架,同时将其应用到其它一些问题,并揭示其与残量型后验误差估计子之间的关系。
中文关键词: 自适应有限元;后验误差估计;收敛性;非局部扩散模型;分数阶索伯列夫空间
英文摘要: Nonlocal properties appear widely in real lives. Therefore, it is very necessary to study the related nonlocal models including peridynamic models and nonlocal diffusion models, which has great practical significance. How to solve nonlocal models fast and efficiently? It is an advanced subject in the field of scientific computing. The adaptive finite element method, one of the most efficient numerical method, is a research hot spot in scientific and engineering computations. This project will mianly study the adaptive finite element algorithms and the corresponding numerical analyses for nonlocal models, which will be very useful to develop high efficient numerical algorithms and softwares for nonlocal models. The main contents of the project are as follows: study the a posteriori error estimation and the convergence of the corresponding adaptive finite element algorithms for nonlocal models with singular kernels; propose a "projecting" a posteriori error estimation method to develop adaptive finite element methods for nonlocal models with hypersingular kernels; establish an abstract framework for "projecting" a posteriori error estimation method with applications in other problems and reveal the relation between the "projecting" error estimator and the well-known residual error estimator.
英文关键词: adaptive finite element;a posteriori error estimation;convergence;nonlocal diffusion model;fractional Sobolev space