1. Lecture 24 Exemplary Inverse Problems including Vibrational Problems

2. Syllabus Lecture 01Describing Inverse Problems Lecture 02Probability and Measurement Error, Part 1 Lecture 03Probability and Measurement Error, Part 2 Lecture 04The L 2 Norm and Simple Least Squares Lecture 05A Priori Information and Weighted Least Squared Lecture 06Resolution and Generalized Inverses Lecture 07Backus-Gilbert Inverse and the Trade Off of Resolution and Variance Lecture 08The Principle of Maximum Likelihood Lecture 09Inexact Theories Lecture 10Nonuniqueness and Localized Averages Lecture 11Vector Spaces and Singular Value Decomposition Lecture 12Equality and Inequality Constraints Lecture 13L 1, L ∞ Norm Problems and Linear Programming Lecture 14Nonlinear Problems: Grid and Monte Carlo Searches Lecture 15Nonlinear Problems: Newton’s Method Lecture 16Nonlinear Problems: Simulated Annealing and Bootstrap Confidence Intervals Lecture 17Factor Analysis Lecture 18Varimax Factors, Empircal Orthogonal Functions Lecture 19Backus-Gilbert Theory for Continuous Problems; Radon’s Problem Lecture 20Linear Operators and Their Adjoints Lecture 21Fréchet Derivatives Lecture 22 Exemplary Inverse Problems, incl. Filter Design Lecture 23 Exemplary Inverse Problems, incl. Earthquake Location Lecture 24 Exemplary Inverse Problems, incl. Vibrational Problems

3. Purpose of the Lecture solve a few exemplary inverse problems tomography vibrational problems determining mean directions

4. Part 1 tomography

5. d i = ∫ ray i m(x(s), y(s)) ds ray i tomography: data is line integral of model function assume ray path is known x y

6. discretization: model function divided up into M pixels m j

7. data kernel G ij = length of ray i in pixel j

8. here’s an easy, approximate way to calculate it

9. ray i start with G set to zero then consider each ray in sequence

10. ∆s divide each ray into segments of arc length ∆s and step from segment to segment

11. determine the pixel index, say j, that the center of each line segment falls within add ∆s to G ij repeat for every segment of every ray

12. You can make this approximation indefinitely accurate simply by decreasing the size of ∆ s (albeit at the expense of increase the computation time)

13. Suppose that there are M=L 2 voxels A ray passes through about L voxels G has NL 2 elements NL of which are non-zero so the fraction of non-zero elements is 1/L hence G is very sparse

14. In a typical tomographic experiment some pixels will be missed entirely and some groups of pixels will be sampled by only one ray

15. the value of these pixels is completely undetermined only the average value of these pixels is determined hence the problem is mixed-determined (and usually M>N as well)

16. so you must introduce some sort of a priori information to achieve a solution say a priori information that the solution is small or a priori information that the solution is smooth

17. Solution Possibilities 1.Damped Least Squares (implements smallness): Matrix G is sparse and very large use bicg() with damped least squares function 2. Weighted Least Squares (implements smoothness): Matrix F consists of G plus second derivative smoothing use bicg() with weighted least squares function

18. Solution Possibilities 1.Damped Least Squares: Matrix G is sparse and very large use bicg() with damped least squares function 2. Weighted Least Squares: Matrix F consists of G plus second derivative smoothing use bicg() with weighted least squares function test case has very good ray coverage, so smoothing probably unnecessary

19. True model x y sources and receivers

20. x y Ray Coverage x y just a “few” rays shown else image is black

21. Data, plotted in Radon-style coordinates angle θ of ray distance r of ray to center of image Lesson from Radon’s Problem: Full data coverage need to achieve exact solution minor data gaps

22. Estimated model x y True model

23. Estimated model x y streaks due to minor data gaps they disappear if ray density is doubled

24. but what if the observational geometry is poor so that broads swaths of rays are missing ?

25. (A) (B) (C) (D) x y x y x y θ r complete angular coverage

26. (A) (B) (C) (D) x y x y x y θ r incomplete angular coverage

27. Part 2 vibrational problems

28. statement of the problem Can you determine the structure of an object just knowing the characteristic frequencies at which it vibrates? frequency

29. the Fréchet derivative of frequency with respect to velocity is usually computed using perturbation theory hence a quick discussion of what that is...

30. perturbation theory a technique for computing an approximate solution to a complicated problem, when 1. The complicated problem is related to a simple problem by a small perturbation 2. The solution of the simple problem must be known

31. simple example

33. we know the solution to this equation: x 0 =±c

37. Here’s the actual vibrational problem acoustic equation with spatially variable sound velocity v

38. acoustic equation with spatially variable sound velocity v frequencies of vibration or eigenfrequencies patterns of vibration or eigenfunctions or modes

39. v(x) = v (0) (x) + εv (1) (x) +... assume velocity can be written as a perturbation around some simple structure v (0) (x)

40. eigenfunctions known to obey orthonormality relationship

41. now represent eigenfrequencies and eigenfunctions as power series in ε

42. represent first-order perturbed shapes as sum of unperturbed shapes now represent eigenfrequencies and eigenfunctions as power series in ε

43. plug series into original differential equation group terms of equal power of ε solve for first-order perturbation in eigenfrequencies ω n (1) and eigenfunction coefficients b nm (use orthonormality in process)

44. result

45. result for eigenfrequencies write as standard inverse problem

46. standard continuous inverse problem

47. perturbation in the eigenfrequencies are the data perturbation in the velocity structure is the model function

48. standard continuous inverse problem depends upon the unperturbed velocity structure, the unperturbed eigenfrequency and the unperturbed mode data kernel or Fréchet derivative

49. 1D organ pipe unperturbed problem has constant velocity 0 h x open end, p=0 closed end dp/dx=0 perturbed problem has variable velocity

50. 0 h x p=0 dp/dx=0 p1p1 xx p2p2 p3p3 x 1 modes frequencies 23 0

51. solution to unperturbed problem

52. position, x velocity, v perturbed unperturbed 00.20.40.60.811.21.41.61.82 0 2 4 6 8 10 velocity structure

53. How to discretize the model function? m is veloctity function evaluated at sequence of points equally spaced in x our choice is very simple

54. the data a list of frequencies of vibration true, unperturbed true, perturbed observed = true, perturbed + noise frequency

55. ωiωi mjmj the data kernel

56. Solution Possibilities 1.Damped Least Squares (implements smallness): Matrix G is not sparse use bicg() with damped least squares function 2. Weighted Least Squares (implements smoothness): Matrix F consists of G plus second derivative smoothing use bicg() with weighted least squares function

57. Solution Possibilities 1.Damped Least Squares (implements smallness): Matrix G is not sparse use bicg() with damped least squares function 2. Weighted Least Squares (implements smoothness): Matrix F consists of G plus second derivative smoothing use bicg() with weighted least squares function our choice

58. position, x velocity, v the solution true estimated

59. position, x velocity, v the solution true estimated

60. mimi mjmj the model resolution matrix

61. mimi mjmj what is this?

62. This problem has a type of nonuniqueness that arises from its symmetry a positive velocity anomaly at one end of the organ pipe trades off with a negative anomaly at the other end

63. this behavior is very common and is why eigenfrequency data are usually supplemented with other data e.g. travel times along rays that are not subject to this nonuniqueness

64. Part 3 determining mean directions

65. statement of the problem you measure a bunch of directions (unit vectors) what’s their mean? x y z data mean

66. what’s a reasonable probability density function for directional data? Gaussian doesn’t quite work because its defined on the wrong interval (- ∞, +∞)

67. θ ϕ central vector datum coordinate system distribution should be symmetric in ϕ

68. angle, θ π-π-πp(θ) Fisher distribution similar in shape to a Gaussian but on a sphere

69. angle, θ π-π-πp(θ) Fisher distribution similar in shape to a Gaussian but on a sphere “precision parameter” quantifies width of p.d.f. κ=5 κ=1

70. solve by direct application of principle of maximum likelihood

71. maximize joint p.d.f. of data with respect to κ and cos(θ)

72. x : Cartesian components of observed unit vectors m : Cartesian components of central unit vector; must constrain |m|=1

73. likelihood function constraint unknowns m, κ C = = 0

74. Lagrange multiplier equations

75. Results valid when κ>5

76. Results central vector is parallel to the vector that you get by putting all the observed unit vectors end-to-end

77. Solution Possibilities Determine m by evaluating simple formula 1.Determine κ using simple but approximate formula 2. Determine κ using bootstrap method our choice only valid when κ>5

78. Application to Subduction Zone Stresses Determine the mean direction of P-axes of deep (300-600 km) earthquakes in the Kurile-Kamchatka subduction zone

79. N E datacentral directionbootstrap