摘要
讨论了广义神经传播方程的低阶H^1-Galerkin混合元方法.其逼近空间不需要满足LBB条件,并且在不需要采用Ritz投影的情况下,通过插值算子,平均值技巧和高精度分析结果得到了超逼近性质,进而通过插值后处理技术导出了H^1-模的整体超收敛结果.
A low-order H^1-Galerkin mixed finite elements method of the generalized nerve conductive equations is proposed, in which the approximating spaces needn't satisfy the LBB consistency condition. Moreove, by use of interpolation operator, mean-value technique and high accucy results instead of the traditional Ritz projection, the superclose properties are obtained. Furthermore, the global superconvergences with H^1-norm are derived through constructing the interpolated post-processing operator.
出处
《数学的实践与认识》
北大核心
2015年第10期229-237,共9页
Mathematics in Practice and Theory
基金
国家自然科学基金(11101381
11271340)