1 function [x, k] = preconditioned_conjugate_gradient_method(Q, ...

2 M, ...

3 b, ...

4 x0, ...

5 tolerance, ...

6 max_iterations)

7 %

8 % Solve,

9 %

10 % Qx = b

11 %

12 % or equivalently,

13 %

14 % min [phi(x) = (1/2)*<Qx,x> + <b,x>]

15 %

16 % using the preconditioned conjugate gradient method (14.56 in

17 % Guler). If ``M`` is the identity matrix, we use the slightly

18 % faster implementation in conjugate_gradient_method.m.

19 %

20 % INPUT:

21 %

22 % - ``Q`` -- The coefficient matrix of the system to solve. Must

23 % be positive definite.

24 %

25 % - ``M`` -- The preconditioning matrix. If the actual matrix used

26 % to precondition ``Q`` is called ``C``, i.e. ``C^(-1) * Q *

27 % C^(-T) == \bar{Q}``, then M=CC^T. However the matrix ``C`` is

28 % never itself needed. This is explained in Guler, section 14.9.

29 %

30 % - ``b`` -- The right-hand-side of the system to solve.

31 %

32 % - ``x0`` -- The starting point for the search.

33 %

34 % - ``tolerance`` -- How close ``Qx`` has to be to ``b`` (in

35 % magnitude) before we stop.

36 %

37 % - ``max_iterations`` -- The maximum number of iterations to

38 % perform.

39 %

40 % OUTPUT:

41 %

42 % - ``x`` - The computed solution to Qx=b.

43 %

44 % - ``k`` - The ending value of k; that is, the number of

45 % iterations that were performed.

46 %

47 % NOTES:

48 %

49 % All vectors are assumed to be *column* vectors.

50 %

51 % The cited algorithm contains a typo; in "The Preconditioned

52 % Conjugate-Gradient Method", we are supposed to define

53 % d_{0} = -z_{0}, not -r_{0} as written.

54 %

55 % The rather verbose name of this function was chosen to avoid

56 % conflicts with other implementations.

57 %

58 % REFERENCES:

59 %

60 % 1. Guler, Osman. Foundations of Optimization. New York, Springer,

61 % 2010.

62 %

63 % 2. Shewchuk, Jonathan Richard. An Introduction to the Conjugate

64 % Gradient Method Without the Agonizing Pain, Edition 1.25.

65 % August 4, 1994.

66 %

68 % We use this in the inner loop.

69 sqrt_n = floor(sqrt(length(x0)));

71 % Set k=0 first, that way the references to xk,rk,zk,dk which

72 % immediately follow correspond (semantically) to x0,r0,z0,d0.

73 k = 0;

75 xk = x0;

76 rk = Q*xk - b;

77 zk = M \ rk;

78 dk = -zk;

80 while (k <= max_iterations)

82 if (norm(rk) < tolerance)

83 % Check our stopping condition. This should catch the k=0 case.

84 x = xk;

85 return;

86 end

88 % Used twice, avoid recomputation.

89 rkzk = rk' * zk;

91 % The term alpha_k*dk appears twice, but so does Q*dk. We can't

92 % do them both, so we precompute the more expensive operation.

93 Qdk = Q * dk;

95 alpha_k = rkzk/(dk' * Qdk);

96 x_next = xk + (alpha_k * dk);

98 % The recursive definition of r_next is prone to accumulate

99 % roundoff error. When sqrt(n) divides k, we recompute the

100 % residual to minimize this error. This modification is due to the

101 % second reference.

102 if (mod(k, sqrt_n) == 0)

103 r_next = Q*x_next - b;

104 else

105 r_next = rk + (alpha_k * Qdk);

106 end

108 z_next = M \ r_next;

109 beta_next = (r_next' * z_next)/rkzk;

110 d_next = -z_next + beta_next*dk;

112 % We potentially just performed one more iteration than necessary

113 % in order to simplify the loop. Note that due to the structure of

114 % our loop, we will have k > max_iterations when we fail to

115 % converge.

116 k = k + 1;

117 xk = x_next;

118 rk = r_next;

119 zk = z_next;

120 dk = d_next;

121 end

123 % The algorithm didn't converge, but we still want to return the

124 % terminal value of xk.

125 x = xk;

126 end