minimise.coordinate

47 - def __init__(self, func, dfunc, args, x0, min_options, func_tol, grad_tol, maxiter, a0, mu, eta, full_output, print_flag, print_prefix):

48 """Class for back-and-forth coordinate descent minimisation specific functions. 49 50 Unless you know what you are doing, you should call the function 'coordinate_descent' rather 51 than using this class. 52 """ 53 54 # Function arguments. 55 self.func = func 56 self.dfunc = dfunc 57 self.args = args 58 self.xk = x0 59 self.func_tol = func_tol 60 self.grad_tol = grad_tol 61 self.maxiter = maxiter 62 self.full_output = full_output 63 self.print_flag = print_flag 64 self.print_prefix = print_prefix 65 66 # Set a0. 67 self.a0 = a0 68 69 # Line search constants for the Wolfe conditions. 70 self.mu = mu 71 self.eta = eta 72 73 # Initialisation failure flag. 74 self.init_failure = 0 75 76 # Setup the line search options and algorithm. 77 self.line_search_options(min_options) 78 self.setup_line_search() 79 80 # Initialise the function, gradient, and Hessian evaluation counters. 81 self.f_count = 0 82 self.g_count = 0 83 self.h_count = 0 84 85 # Initialise the warning string. 86 self.warning = None 87 88 # Set the convergence test function. 89 self.setup_conv_tests() 90 91 # The initial function value and gradient vector. 92 self.fk, self.f_count = apply(self.func, (self.xk,)+self.args), self.f_count + 1 93 self.dfk, self.g_count = apply(self.dfunc, (self.xk,)+self.args), self.g_count + 1 94 95 # Create the coordinate descent directions, and initialise the coordinate descent iteration number and direction flag. 96 self.cd_dir = identity(len(self.xk), Float64) 97 self.n = 0 98 self.back = 0

99 100

101 - def new_param_func(self):

102 """The new parameter function. 103 104 Find the search direction, do a line search, and get xk+1 and fk+1. 105 """ 106 107 # Get the coordinate descent direction (pk is forced to be a descent direction). 108 if dot(self.dfk, self.cd_dir[self.n]) > 0.0: 109 self.pk = -self.cd_dir[self.n] 110 else: 111 self.pk = self.cd_dir[self.n] 112 113 # Line search. 114 self.line_search() 115 116 # Find the new parameter vector and function value at that point. 117 self.xk_new = self.xk + self.alpha * self.pk 118 self.fk_new, self.f_count = apply(self.func, (self.xk_new,)+self.args), self.f_count + 1 119 self.dfk_new, self.g_count = apply(self.dfunc, (self.xk_new,)+self.args), self.g_count + 1 120 121 # Scale the coordinate direction to minimise the number of function calls. 122 self.cd_dir[self.n] = self.alpha * self.pk

123 124

125 - def update(self):

126 """Function to update the function value, gradient vector, and Hessian matrix.""" 127 128 # Update the coordinate descent iteration number and direction flag. 129 if not self.back: 130 if self.n < len(self.xk) - 1: 131 self.n = self.n + 1 132 else: 133 self.back = 1 134 self.n = self.n - 1 135 else: 136 if self.n > 0: 137 self.n = self.n - 1 138 else: 139 self.back = 0 140 self.n = self.n + 1 141 if self.print_flag >= 2: 142 print self.print_prefix + "back_flag: " + `self.back` 143 print self.print_prefix + "n: " + `self.n` 144 145 # Store old data. 146 self.fk_last = self.fk 147 self.dfk_last = self.dfk * 1.0 148 149 # Shift k+1 data to k. 150 self.xk = self.xk_new * 1.0 151 self.fk = self.fk_new 152 self.dfk = self.dfk_new * 1.0

Source Code for Module minimise.coordinate_descent