Download presentation
Presentation is loading. Please wait.
Published byMaurice Mathews Modified over 9 years ago
1
A Localized Method of Particular Solutions for Solving Near Singular Problems C.S. Chen, Guangming Yao, D.L. Young Department of Mathematics University of Southern Mississippi U.S.A.
2
2016/1/142 OutlineOutline Radial Basis Functions The global approaches of the method of particular solutions Numerical examples of global method Local approach of the method of particular solutions Numerical examples of local method Near Singular Problems
3
2016/1/143 Radial Basis Functions Linear: Cubic: Multiquadrics: Polyharmonic Spines: Gaussian:
4
2016/1/144 Assume that To approximate f bywe usually require fitting the given data set of pairwise distinct centres with the imposed conditions The linear system is well-posed if the interpolation matrix is non-singular Surface Reconstruction Scheme
5
2016/1/145 The Splitting Method Consider the following equation Whereis a bounded open nonempty domain with sufficiently regular boundary Letwheresatisfying but does not necessary satisfy the boundary condition in (11). (10) (11) v satisfies (12) (13) (14)
6
2016/1/146 Assume that and that we can obtain an analytical solutionto Then To approximate f bywe usually require fitting the given data set of pairwise distinct centres with the imposed conditions Particular Solutions
7
2016/1/147 The linear system is well-posed if the interpolation matrix is non-singular where and (*)
8
2016/1/148 For in 2D
9
2016/1/149
10
10
11
2016/1/1411 Where G(r) is the fundamental solution of L Boundary Method is required.
12
2016/1/1412 The Method of Particular Solutions (MPS) where
13
2016/1/1413 Impose boundary conditions
14
2016/1/1414
15
2016/1/1415
16
2016/1/1416 Numerical Results
17
2016/1/1417 Example I Analytical solution: Computational Domain:
18
2016/1/1418
19
2016/1/1419 c : shape parameter of MQ
20
2016/1/1420 Consider the Poisson’s equation Given a large data set where
21
2016/1/1421
22
2016/1/1422
23
2016/1/1423
24
2016/1/1424
25
2016/1/1425
26
2016/1/1426
28
2016/1/1428
29
2016/1/1429
30
2016/1/1430
31
2016/1/1431
32
2016/1/14 Non-Dirichlet boundary condition
33
2016/1/1433
34
2016/1/1434
35
2016/1/1435
36
2016/1/1436
37
2016/1/1437
38
2016/1/1438
39
2016/1/1439 The absolute errors of LMAPS with L=1, n=5, S n =100, c=8.9
40
2016/1/14 L=1, S n = 100, N=225.
41
2016/1/1441 Local MPS verse Global MPS
42
2016/1/1442 n: number of neighbor points
43
2016/1/1443
44
2016/1/1444
45
2016/1/1445 LMPS verse LMQ
46
2016/1/14 Near Singular Problem I C.S. Chen, G. Kuhn, J. Li, G. Mishuris, Radial basis functions for solving near singular Poisson’s problems, Communication in Numerical Methods in Engineering, 2003, 19, 333-347.
47
2016/1/14
48
48 Profile of exact solution
49
2016/1/1449 CS-RBF 400 quasi-random points
50
2016/1/1450 Test 1Test 2
51
2016/1/1451 Normalized Shape parameter where
52
2016/1/1452
53
2016/1/1453 Sobel quasi-random nodes Von-Del Corput quasi-random nodes Random nodes
54
2016/1/1454 Speed up N=10,000 CPU = 0.5/3.42 s N=40,000 CPU = 3.31/14.06 s N=62,500 CPU = 7.01/25.28 s
55
2016/1/1455 RMSE error verse shape parameter for a=1.6 and various mesh sizes LMPS
56
2016/1/1456 RMSE error verse shape parameter for h=1/200, and various value of a.
57
2016/1/1457
58
2016/1/1458 Near Singular Problem II Exact solution
59
2016/1/14 Profile of f(x,y) f(1,1,) = -15,861, f(0,0)=237
60
2016/1/1460
61
2016/1/1461 Near Singular Problem III
62
2016/1/1462
63
2016/1/1463 Adaptive Method First step Second step
64
2016/1/1464 3 rd step 4 th step
65
2016/1/1465
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.