Module: EvoSynth::Problems::FloatBenchmarkFuntions

Defined in:
lib/evosynth/problems/float_benchmark_functions.rb

Overview

This module contains some multi-dimensional Benchmarkfunctions. You simply set the number of dimensions by the length of the given (Float) array.

Class Method Summary collapse

Class Method Details

.ackley(xs) ⇒ Object

Ackley (Ackley 1987)

global minimum: f(x) = 0 at x(i) = 0, i = 1..n



81
82
83
84
85
 
# File 'lib/evosynth/problems/float_benchmark_functions.rb', line 81

def FloatBenchmarkFuntions.ackley(xs)
	quad_sum = xs.inject(0.0) { |sum, x| sum += x**2 }
	cos_sum = xs.inject(0.0) { |sum, x| sum += Math.cos(2 * Math::PI * x) }
	20 + Math::E - 20 * Math.exp(-0.2 * Math.sqrt( (1.0 / xs.size) * quad_sum )) - Math.exp( (1.0 / xs.size) * cos_sum )
end

.double_sum(xs) ⇒ Object

Double Sum function (Schwefel 1977)

global minimum: f(x) = 0 at x(i) = 0, i = 1..n



45
46
47
48
49
 
# File 'lib/evosynth/problems/float_benchmark_functions.rb', line 45

def FloatBenchmarkFuntions.double_sum(xs)
	xs.inject(0.0) do |sum, i|
		sum += xs[0..i].inject(0.0) { |sum, x| sum += x }
	end
end

.griewank(xs) ⇒ Object

Griewank (Törn & Zilinskas 1989)

global minimum: f(x) = 0 at x(i) = 0, i = 1..n



91
92
93
94
95
 
# File 'lib/evosynth/problems/float_benchmark_functions.rb', line 91

def FloatBenchmarkFuntions.griewank(xs)
	product = 1.0
	xs.each_with_index { |x, i| product *= Math.cos x / Math.sqrt(i+1) }
	1 + xs.inject(0.0) { |sum, x| sum += x**2 / (400 * xs.size) } - product
end

.rastgrin(xs) ⇒ Object

Rastgrin (Törn & Zilinskas 1989)

global minimum: f(x) = 0 at x(i) = 0, i = 1..n



63
64
65
 
# File 'lib/evosynth/problems/float_benchmark_functions.rb', line 63

def FloatBenchmarkFuntions.rastgrin(xs)
	10 * xs.size + xs.inject(0.0) { |sum, x| sum += x**2 - 10 * Math.cos(2 * Math::PI * x) }
end

.rosenbrock(xs) ⇒ Object

Rosenbrock (De Jong 1975)

global minimum: f(x) = 0 at x(i) = 1, i = 1..n



71
72
73
74
75
 
# File 'lib/evosynth/problems/float_benchmark_functions.rb', line 71

def FloatBenchmarkFuntions.rosenbrock(xs)
	(0..xs.size - 2).inject(0.0) do |sum, i|
		sum += 100 * (xs[i]**2 - xs[i+1])**2 + (1 - xs[i])**2
	end
end

.sinus_sum(xs) ⇒ Object

Sinus Sum function (Schwefel 1995)

global minimum: f(x) = 0 at x(i) = 420.9687, i = 1..n



37
38
39
 
# File 'lib/evosynth/problems/float_benchmark_functions.rb', line 37

def FloatBenchmarkFuntions.sinus_sum(xs)
	418.98289 * xs.size + xs.inject(0.0) { |sum, x| sum += -x * Math.sin(Math.sqrt(x.abs)) }
end

.sphere(xs) ⇒ Object

Sphere function (Rechenberg 1973 and De Jong 1975)

global minimum: f(x) = 0 at x(i) = 0, i = 1..n



55
56
57
 
# File 'lib/evosynth/problems/float_benchmark_functions.rb', line 55

def FloatBenchmarkFuntions.sphere(xs)
	xs.inject(0.0) { |sum, x| sum += x**2 }
end