##
**\(H\)-\(p\) clouds – an \(h\)-\(p\) meshless method.**
*(English)*
Zbl 0869.65069

The so-called meshless methods (unlike the finite element or the finite difference methods) can dramatically simplify the large scale simulation of physical phenomena. Among the first methods of these types have been the methods due to T. Liszka [Int. J. Numer. Methods Eng. 20, 1599-1612 (1984; Zbl 0544.65006)] or to T. Belytschko, Y. Y. Lu, L. Gu [ibid. 37, No. 2, 229-256 (1994; Zbl 0796.73077)], based on ideas of P. Lancaster, K. Salkauskas [Math. Comp. 37, 141-158 (1981; Zbl 0469.41005)].

The article presents a new family of meshless methods for the solution of boundary value problems. The method uses radial basis functions of varying size of supports. Namely, the definition of a class of functions, to be used as trial and test functions for Galerkin approximations, is based on a partition of unity, connected with a scattered set of nodes. The idea behind that definition is the hierarchical addition of appropriate elements to the partition functions, that the resulting set represents (through linear combinations) polynomials of certain higher degree.

- Numerical experiments with the new technique in one or two dimensions show very promising results: Evidently the recent paper is a part of a running discussion. One of its essential aims is to bring more stringency (and this means: more analysis) into that domain of numerical mathematics.

The article presents a new family of meshless methods for the solution of boundary value problems. The method uses radial basis functions of varying size of supports. Namely, the definition of a class of functions, to be used as trial and test functions for Galerkin approximations, is based on a partition of unity, connected with a scattered set of nodes. The idea behind that definition is the hierarchical addition of appropriate elements to the partition functions, that the resulting set represents (through linear combinations) polynomials of certain higher degree.

- Numerical experiments with the new technique in one or two dimensions show very promising results: Evidently the recent paper is a part of a running discussion. One of its essential aims is to bring more stringency (and this means: more analysis) into that domain of numerical mathematics.

Reviewer: E.Lanckau (Chemnitz)

### MSC:

65N30 | Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs |

35J20 | Variational methods for second-order elliptic equations |

### Keywords:

finite methods; Galerkin methods; numerical experiments; meshless methods; radial basis functions
PDF
BibTeX
XML
Cite

\textit{C. A. Duarte} and \textit{J. T. Oden}, Numer. Methods Partial Differ. Equations 12, No. 6, 673--705 (1996; Zbl 0869.65069)

Full Text:
DOI

### References:

[1] | , and , Crack propagation by element free Galerkin methods, in Advanced Computational Methods for Material Modeling, p. 191, 1993, AMD-Vol. 180/PVP-Vol. 268, ASME 1993. |

[2] | Liszka, Inter. J. Numer. Meth. Eng. 20 pp 1599– (1984) · Zbl 0544.65006 |

[3] | and , Finite difference method for arbitrary irregular meshes in nonlinear problems of applied mechanics, in IV SMiRt, San Francisco, 1977. |

[4] | Liszka, Comput. Struct. 11 pp 83– (1980) · Zbl 0427.73077 |

[5] | Gingold, J. Comp. Phys. 46 pp 429– (1982) · Zbl 0487.76010 |

[6] | Lucy, Astron. J. 82 pp 1013– (1977) |

[7] | Monaghan, SIAM J. Sci. Stat. Comput. 3 pp 422– (1982) · Zbl 0498.76010 |

[8] | Monaghan, Comp. Phys. Comm. 48 pp 89– (1988) · Zbl 0673.76089 |

[9] | Nayroles, Comp. Mech. 10 pp 307– (1992) · Zbl 0764.65068 |

[10] | Belytschko, Int. J. Num. Meth. Eng. 37 pp 229– (1994) · Zbl 0796.73077 |

[11] | Amarantuga, Int. J. Num. Meth. Eng. 37 pp 2703– (1994) · Zbl 0813.65106 |

[12] | , , and , Wavelet solutions of linear and nonlinear elliptic, parabolic and hyperbolic problems in one space dimension, in ”9th Int. Conf. on Numerical Methods in Applied Sciences and Engineering,” SIAM, Philadelphia, 1990. |

[13] | Qian, J. Comp. Phys. 106 pp 155– (1993) · Zbl 0771.65072 |

[14] | Kansa, Comput. Math. Appl. 19 pp 127– (1990) · Zbl 0692.76003 |

[15] | Kansa, Comput. Math. Appl. 19 pp 147– (1990) · Zbl 0850.76048 |

[16] | Liu, Inter. J. Num. Meth. Fluids |

[17] | Kam Liu, Inter. J. Num. Meth. Eng. |

[18] | Some remarks on free mesh method: A kind of meshless finite element method, in International Conference on Computational Engineering Science, Hawaii, August 1995. |

[19] | Lancaster, Math. Comp. 37 pp 141– (1981) |

[20] | and , Curve and Surface Fitting, an Introduction, Academic Press, San Diego, 1986. · Zbl 0649.65012 |

[21] | A review of some meshless methods to solve partial differential equations, Technical Report 95-06, TICAM, The University of Texas at Austin, 1995. |

[22] | , , , and , An analysis of the smoothed particle hydrodynamics, Technical Report SAND93-2513 UC-705, Sandia, 1994. |

[23] | Lu, Comput. Meth. Appl. Mech. Eng. |

[24] | Franke, Math. Comp. 38 pp 181– (1982) |

[25] | Lu, Comput. Meth. Appl. Mech. Eng. 113 pp 397– (1994) · Zbl 0847.73064 |

[26] | Personal communication, 1994. |

[27] | and , Hp clouds-a meshless method to solve boundary-value problems Technical Report 95-05, TICAM, The University of Texas at Austin, 1995. |

[28] | Sobolev Spaces. Academic Press, Boston, 1978. |

[29] | A two-dimensional function for irregularly spaced data, in ACM Nat. Conf. p. 517, 1968. |

[30] | and , Nodal integration of the element-free Galerkin method, pre-print, 1995. |

[31] | The Finite Element Method for Elliptic Problems, North-Holland, Amsterdam, 1978. |

[32] | and , Texas Finite Element Series Volume IV-Mathematical Aspects, Prentice-Hall, New Jersey, 1983. |

[33] | Babuska, SIAM J. Num. Anal. 15 pp 736– (1978) · Zbl 0398.65069 |

[34] | ”On Generalized Finite Element Methods,” Ph. D. thesis, The University of Maryland, 1995. |

This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.