##
**Circumventing the ill-conditioning problem with multiquadric radial basis functions: Applications to elliptic partial differential equations.**
*(English)*
Zbl 0955.65086

Summary: W. R. Madych and S. A. Nelson [Math. Comput. 54, No. 189, 211-230 (1990; Zbl 0859.41004)] proved multiquadric (MQ) mesh-independent radial basis functions (RBFs) enjoy exponential convergence. The primary disadvantage of the MQ scheme is that it is global, hence, the coefficient matrices obtained from this discretization scheme are full. Full matrices tend to become progressively more ill-conditioned as the rank increases.

In this paper, we explore several techniques, each of which improves the conditioning of the coefficient matrix and the solution accuracy. The methods that were investigated are

(1) replacement of global solvers by block partitioning, LU decomposition schemes,

(2) matrix preconditioners,

(3) variable MQ shape parameters based upon the local radius of curvature of the function being solved,

(4) a truncated MQ basis function having a finite, rather than a full band-width,

(5) multizone methods for large simulation problems, and

(6) knot adaptivity that minimizes the total number of knots required in a simulation problem.

The hybrid combination of these methods contribute to very accurate solutions.

Even though the finite element method (FEM) gives rise to sparse coefficient matrices, these matrices in practice can become very ill-conditioned. We recommend using what has been learned from the FEM practitioners and combining their methods with what has been learned in RBF simulations to form a flexible, hybrid approach to solve complex multidimensional problems.

In this paper, we explore several techniques, each of which improves the conditioning of the coefficient matrix and the solution accuracy. The methods that were investigated are

(1) replacement of global solvers by block partitioning, LU decomposition schemes,

(2) matrix preconditioners,

(3) variable MQ shape parameters based upon the local radius of curvature of the function being solved,

(4) a truncated MQ basis function having a finite, rather than a full band-width,

(5) multizone methods for large simulation problems, and

(6) knot adaptivity that minimizes the total number of knots required in a simulation problem.

The hybrid combination of these methods contribute to very accurate solutions.

Even though the finite element method (FEM) gives rise to sparse coefficient matrices, these matrices in practice can become very ill-conditioned. We recommend using what has been learned from the FEM practitioners and combining their methods with what has been learned in RBF simulations to form a flexible, hybrid approach to solve complex multidimensional problems.

### MSC:

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

65N55 | Multigrid methods; domain decomposition for boundary value problems involving PDEs |

65H10 | Numerical computation of solutions to systems of equations |

65F35 | Numerical computation of matrix norms, conditioning, scaling |

35J65 | Nonlinear boundary value problems for linear elliptic equations |

65N12 | Stability and convergence of numerical methods for boundary value problems involving PDEs |

### Keywords:

ill-conditioned matrices; nonlinear Poisson equation; numerical examples; multiquadric radial basis functions; domain decomposition methods; multizone decomposition methods; exponential convergence; matrix preconditioners; finite element method; sparse coefficient matrices### Citations:

Zbl 0859.41004
PDF
BibTeX
XML
Cite

\textit{E. J. Kansa} and \textit{Y. C. Hon}, Comput. Math. Appl. 39, No. 7--8, 123--137 (2000; Zbl 0955.65086)

Full Text:
DOI

### References:

[1] | Madych, W.R.; Nelson, S.A., Multivariable interpolation and conditionally positive definite functions II, Math. comp., 54, 211-230, (1990) · Zbl 0859.41004 |

[2] | Franke, R., Scattered data interpolation: tests of some methods, Math. comput., 38, 181-199, (1972) |

[3] | Hardy, R.L., Multiquadric equations of topography and other irregular surfaces, J. geophys. res., 76, 1905-1915, (1971) |

[4] | Hardy, R.L., Theory and applications of the multiquadric-biharmonic method: 20 years of discovery, Computers math. applic., 19, 8/9, 163-208, (1990) · Zbl 0692.65003 |

[5] | Duchon, J., Spline minimizing rotation-invariant seminorms in Sobolev space, () |

[6] | Kansa, E.J., Multiquadrics—A scattered data approximation scheme with applications to computational fluid dynamics—I. surface approximations and partial derivative estimates, Computers math. applic., 19, 8/9, 127-145, (1990) · Zbl 0692.76003 |

[7] | Kansa, E.J., Multiquadrics—A scattered data approximation scheme with applications to computational fluid dynamics—II. solutions to hyperbolic, parabolic, and elliptic partial differential equations, Computers math. applic., 19, 8/9, 147-161, (1990) · Zbl 0850.76048 |

[8] | Golberg, M.A.; Chen, C.S., The theory of radial basis functions used in the dual reciprocity boundary element method, Appl. math. comp., 60, 125-136, (1994) |

[9] | Sharan, M.; Kansa, E.J.; Gupta, S., Applications of the multiquadric method for the solution of elliptic partial differential equations, Appl. math. comput., 84, 275-302, (1997) · Zbl 0883.65083 |

[10] | Golberg, M.A.; Chen, C.S., Discrete projection methods for integral equations, (1997), Comput. Mech Boston, MA · Zbl 0903.76065 |

[11] | Hon, Y.C.; Mao, X.Z., A multiquadric interpolation method for solving initial value problems, Sci. comput., 12, 1, 51-55, (1997) · Zbl 0907.65062 |

[12] | Hon, Y.C.; Lu, M.W.; Xue, W.M.; Zhu, Y.M., Multiquadric method for the numerical solution of a biphasic model, Appl. math. comput., 88, 153-175, (1997) · Zbl 0910.76059 |

[13] | Hon, Y.C.; Mao, X.Z., An efficient numerical scheme for Burgers equation, Appl. math. comput., 95, 37-50, (1998) · Zbl 0943.65101 |

[14] | Hon, Y.C.; Cheung, K.F.; Mao, X.Z.; Kansa, E.J., A multiquadric solution for shallow water equation, ASCE J. hydraulic engineering, 125, 5, 524-533, (1999) |

[15] | Hon, Y.C.; Mao, X.Z., A radial basis function method for solving options pricing model, Financial engineering, 8, 1, 31-50, (1999) · Zbl 0943.65101 |

[16] | Wu, Z.M., Hermite-bikhoff interpolation of scattered data by radial basis function, Approximation theory and its application, 8, 1-10, (1992) |

[17] | Wu, Z.M., Solving PDE with radial basis function and the error estimation, (), GuangZhou |

[18] | Dubal, M.R.; Olivera, S.R.; Matzner, R.A., () |

[19] | Schaback, R.; Hon, Y.C., On unsymmetric collocation by radial basis functions, Appl. math. comput., (1999), (to appear) · Zbl 1026.65107 |

[20] | Smith, B.; Bjorstad, P.; Gropp, W., Domain decomposition: parallel multilevel methods for elliptic partial differential equations, (1996), Cambridge University Press Cambridge, UK · Zbl 0857.65126 |

[21] | Demmel, J.W.; Higham, N.J.; Schreiber, R.S., Stability of block LU factorization, Numer. lin. algebra applic., 2, 2, 173-190, (1995) · Zbl 0834.65010 |

[22] | Dubal, M.R., Domain decomposition and local refinement for multiquadric approximations. I. second-order equations in one-dimension, Int. J. appl. sci. comput., 1, 1, 146-171, (1994) |

[23] | Madych, W.R., Miscellaneous error bounds for multiquadric and related interpolants, Computers math. applic., 24, 12, 121-138, (1992) · Zbl 0766.41003 |

[24] | Tarwater, A.E., A parametric study of Hardy’s multiquadric equations for scattered data interpolation, (1985), UCRL-54670 |

[25] | Carlson, R.E.; Foley, T.A., The parameter R2 in multiquadric interpolation, Computers math. applic., 21, 9, 29-42, (1991) · Zbl 0725.65009 |

[26] | Milroy, M.F.; Vichers, G.W.; Bradley, C., An adaptive radial basis function approach to modeling scattered data, J. appl. sci. and comput., 1, 2, 319-349, (1994) |

[27] | Golberg, M.A.; Chen, C.S., Improved multiquadric approximation for partial differential equations, Engr. anal. with bound. elem., 18, 9-17, (1996) |

[28] | Hickernell, F.J.; Hon, Y.C., Radial basis function approximation of the surface wind field from scattered data, Appl. sci. comput., 4, 221-247, (1998) |

[29] | Chen, C.S.; Brebbia, C.A.; Power, H., Boundary element methods using compactly supported radial basis functions, Comm. numer. meth. engng, 15, 137-150, (1999) · Zbl 0927.65140 |

[30] | Wendland, H., Piecewise polynomial, positive definite and compactly supported radial basis functions of minimal degree, Adv. comput. math., 4, 389-396, (1995) · Zbl 0838.41014 |

[31] | Wong, S.M.; Hon, Y.C.; Li, T.S.; Chung, S.L.; Kansa, E.J., Multizone decomposition for simulation of time-dependent problems using the multiquadric scheme, Computers math. applic., 37, 8, 23-43, (1999) · Zbl 0951.76066 |

[32] | Fasshauer, G.; Jerome, J., Multistep approximations algorithms: improved convergence rates through postconditioning with smoothing kernals, Adv. comp. math., 10, 1-27, (1999) · Zbl 0924.65009 |

[33] | Micchelli, C.A., Interpolation of scattered data: distance matrices and conditionally positive definite functions, Constr. approx., 2, 11-22, (1986) · Zbl 0625.41005 |

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.