Class: RV::BoxMuller

Inherits:
Object
  • Object
show all
Includes:
RV_Generator
Defined in:
lib/random_variates.rb

Overview

Alternate normal/Gaussian random variate generator with specified mean and standard deviation. Defaults to a standard normal.

Arguments

  • mu -> the expected value (default: 0).

  • sigma -> the standard deviation (default: 1).

  • rng -> the (Enumerable) source of U(0, 1)‘s (default: U_GENERATOR)

Constant Summary collapse

TWO_PI = 
2.0 * Math::PI

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from RV_Generator

#each

Constructor Details

#initialize(mu: 0.0, sigma: 1.0, rng: U_GENERATOR) ⇒ BoxMuller

Returns a new instance of BoxMuller.



234
235
236
237
238
239
240
241
242
 
# File 'lib/random_variates.rb', line 234

def initialize(mu: 0.0, sigma: 1.0, rng: U_GENERATOR)
  raise 'Standard deviation must be positive.' if sigma <= 0

  @mu = mu
  @sigma = sigma
  @rng = rng
  @need_new_pair = false
  # @next_norm = 0.0
end

Instance Attribute Details

#mu ⇒ Object

Returns the value of attribute mu.



232
233
234
 
# File 'lib/random_variates.rb', line 232

def mu
  @mu
end

#sigma ⇒ Object

Returns the value of attribute sigma.



232
233
234
 
# File 'lib/random_variates.rb', line 232

def sigma
  @sigma
end

Instance Method Details

#next ⇒ Object 



244
245
246
247
248
249
250
251
252
253
 
# File 'lib/random_variates.rb', line 244

def next
  if @need_new_pair ^= true
    theta = TWO_PI * @rng.next
    d = @sigma * Math.sqrt(-2.0 * Math.log(@rng.next))
    @next_norm = @mu + d * Math.sin(theta)
    @mu + d * Math.cos(theta)
  else
    @next_norm
  end
end