摘要
利用径向基函数在Sobolev空间Hk(Ω) (k >n2 )中的插值性质 ,由一类特殊的径向函数构成H1 (Ω)空间中的一组基 ,得到求解偏微分方程边值问题的无网格算法 ,并针对散乱数据的特点 ,给出计算整体稠密度h的算法及如何通过加密节点使h值缩小的一个可行的方法 。
By means of the interpolation property of radial basis functions in H^k(Ω) (k>n/2),a meshless method for solving partial differential equations is derived from a basis of H^1(Ω),which is constructed with a kind of special radial functions.An algorithm of computing the value of global data density h for scattered notes and an effective method to reduce the value of h by increasing the number of notes are also given.Finally some numerical expriments are presented by using Sobolev splines and compactly supported positive definite radial basis functions.
出处
《复旦学报(自然科学版)》
CAS
CSCD
北大核心
2004年第3期292-299,共8页
Journal of Fudan University:Natural Science