Class: Integration

Inherits:
Object
  • Object
show all
Defined in:
lib/integration.rb,
lib/integration/version.rb,
lib/integration/methods.rb

Overview

Diverse integration methods Use Integration.integrate as wrapper to direct access to methods

Constant Summary

MInfinity =

Minus Infinity

:minfinity
Infinity =

Infinity

:infinity
RUBY_METHODS =

Pure Ruby methods available.

[:rectangle, :trapezoid, :simpson, :adaptive_quadrature,
:gauss, :romberg, :monte_carlo, :gauss_kronrod,
:simpson3by8, :boole, :open_trapezoid, :milne]
GSL_METHODS =

Methods available when using the rb-gsl gem.

[:qng, :qag]
VERSION =
'0.1.4'

Class Method Summary collapse

Class Method Details

.adaptive_quadrature(a, b, tolerance) ⇒ Object

Adaptive Quadrature Calls the Simpson's rule recursively on subintervals in case the error exceeds the desired tolerance +tolerance+ is the desired tolerance of error



98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
# File 'lib/integration/methods.rb', line 98

def adaptive_quadrature(a, b, tolerance)
  h = (b.to_f - a) / 2
  fa = yield(a)
  fc = yield(a + h)
  fb = yield(b)
  s = h * (fa + (4 * fc) + fb) / 3

  helper = proc do |_a, _b, _fa, _fb, _fc, _h, _s, level|
    if level < 1 / tolerance.to_f
      fd = yield(_a + (_h / 2))
      fe = yield(_a + (3 * (_h / 2)))
      s1 = _h * (_fa + (4.0 * fd) + _fc) / 6
      s2 = _h * (_fc + (4.0 * fe) + _fb) / 6
      if ((s1 + s2) - _s).abs <= tolerance
        s1 + s2
      else
        helper.call(_a, _a + _h, _fa, _fc, fd, _h / 2, s1, level + 1) +
        helper.call(_a + _h, _b, _fc, _fb, fe, _h / 2, s2, level + 1)
      end
    else
      fail 'Integral did not converge'
    end

  end
  helper.call(a, b, fa, fb, fc, h, s, 1)
end

.boole(t1, t2, n, &f) ⇒ Object

Boole's Rule +n+ implies number of subdivisions Source: Weisstein, Eric W. "Boole's Rule." From MathWorld—A Wolfram Web Resource



58
59
60
61
62
63
64
65
# File 'lib/integration/methods.rb', line 58

def boole(t1, t2, n, &f)
  d = (t2 - t1) / n.to_f
  ac = 0
  (0..n - 1).each do |i|
    ac += (d / 90.0) * (7 * f[t1 + i * d] + 32 * f[t1 + i * d + d / 4] + 12 * f[t1 + i * d + d / 2] + 32 * f[t1 + i * d + 3 * d / 4] + 7 * f[t1 + (i + 1) * d])
  end
  ac
end

.gauss(t1, t2, n) ⇒ Object

Gaussian Quadrature n-point Gaussian quadrature rule gives an exact result for polynomials of degree 2n − 1 or less



127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
# File 'lib/integration/methods.rb', line 127

def gauss(t1, t2, n)
  case n
    when 1
      z = [0.0]
      w = [2.0]
    when 2
      z = [-0.57735026919, 0.57735026919]
      w = [1.0, 1.0]
    when 3
      z = [-0.774596669241, 0.0, 0.774596669241]
      w = [0.555555555556, 0.888888888889, 0.555555555556]
    when 4
      z = [-0.861136311594, -0.339981043585, 0.339981043585, 0.861136311594]
      w = [0.347854845137, 0.652145154863, 0.652145154863, 0.347854845137]
    when 5
      z = [-0.906179845939, -0.538469310106, 0.0, 0.538469310106, 0.906179845939]
      w = [0.236926885056, 0.478628670499, 0.568888888889, 0.478628670499, 0.236926885056]
    when 6
      z = [-0.932469514203, -0.661209386466, -0.238619186083, 0.238619186083, 0.661209386466, 0.932469514203]
      w = [0.171324492379, 0.360761573048, 0.467913934573, 0.467913934573, 0.360761573048, 0.171324492379]
    when 7
      z = [-0.949107912343, -0.741531185599, -0.405845151377, 0.0, 0.405845151377, 0.741531185599, 0.949107912343]
      w = [0.129484966169, 0.279705391489, 0.381830050505, 0.417959183673, 0.381830050505, 0.279705391489, 0.129484966169]
    when 8
      z = [-0.960289856498, -0.796666477414, -0.525532409916, -0.183434642496, 0.183434642496, 0.525532409916, 0.796666477414, 0.960289856498]
      w = [0.10122853629, 0.222381034453, 0.313706645878, 0.362683783378, 0.362683783378, 0.313706645878, 0.222381034453, 0.10122853629]
    when 9
      z = [-0.968160239508, -0.836031107327, -0.613371432701, -0.324253423404, 0.0, 0.324253423404, 0.613371432701, 0.836031107327, 0.968160239508]
      w = [0.0812743883616, 0.180648160695, 0.260610696403, 0.31234707704, 0.330239355001, 0.31234707704, 0.260610696403, 0.180648160695, 0.0812743883616]
    when 10
      z = [-0.973906528517, -0.865063366689, -0.679409568299, -0.433395394129, -0.148874338982, 0.148874338982, 0.433395394129, 0.679409568299, 0.865063366689, 0.973906528517]
      w = [0.0666713443087, 0.149451349151, 0.219086362516, 0.26926671931, 0.295524224715, 0.295524224715, 0.26926671931, 0.219086362516, 0.149451349151, 0.0666713443087]
    else
      fail "Invalid number of spaced abscissas #{n}, should be 1-10"
  end

  sum = 0
  (0...n).each do |i|
    t = ((t1.to_f + t2) / 2.0) + (((t2 - t1) / 2.0) * z[i])
    sum += w[i] * yield(t)
  end
  ((t2 - t1) / 2.0) * sum
end

.gauss_kronrod(t1, t2, n, points) ⇒ Object

Gauss Kronrod Rule: Provides a 3n+1 order estimate while re-using the function values of a lower-order(n order) estimate Source: "Gauss–Kronrod quadrature formula", Encyclopedia of Mathematics, Springer, ISBN 978-1-55608-010-4



175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
# File 'lib/integration/methods.rb', line 175

def gauss_kronrod(t1, t2, n, points)
  # g7k15
  case points
    when 15

      z = [-0.9914553711208126, -0.9491079123427585, -0.8648644233597691,
           -0.7415311855993945, -0.5860872354676911, -0.4058451513773972,
           -0.20778495500789848, 0.0, 0.20778495500789848,
           0.4058451513773972, 0.5860872354676911, 0.7415311855993945,
           0.8648644233597691, 0.9491079123427585, 0.9914553711208126]

      w = [0.022935322010529224, 0.06309209262997856, 0.10479001032225019,
           0.14065325971552592, 0.1690047266392679, 0.19035057806478542,
           0.20443294007529889, 0.20948214108472782, 0.20443294007529889,
           0.19035057806478542, 0.1690047266392679, 0.14065325971552592,
           0.10479001032225019, 0.06309209262997856, 0.022935322010529224]

    when 21
      # g10k21

      z = [-0.9956571630258081, -0.9739065285171717, -0.9301574913557082,
           -0.8650633666889845, -0.7808177265864169, -0.6794095682990244,
           -0.5627571346686047, -0.4333953941292472, -0.2943928627014602,
           -0.14887433898163122, 0.0, 0.14887433898163122,
           0.2943928627014602, 0.4333953941292472, 0.5627571346686047,
           0.6794095682990244, 0.7808177265864169, 0.8650633666889845,
           0.9301574913557082, 0.9739065285171717, 0.9956571630258081]

      w = [0.011694638867371874, 0.032558162307964725,
           0.054755896574351995, 0.07503967481091996, 0.0931254545836976,
           0.10938715880229764, 0.12349197626206584, 0.13470921731147334,
           0.14277593857706009, 0.14773910490133849, 0.1494455540029169,
           0.14773910490133849, 0.14277593857706009, 0.13470921731147334,
           0.12349197626206584, 0.10938715880229764, 0.0931254545836976,
           0.07503967481091996, 0.054755896574351995, 0.032558162307964725,
           0.011694638867371874]

    when 31
      # g15k31

      z = [-0.9980022986933971, -0.9879925180204854, -0.9677390756791391,
           -0.937273392400706, -0.8972645323440819, -0.8482065834104272,
           -0.790418501442466, -0.7244177313601701, -0.650996741297417,
           -0.5709721726085388, -0.4850818636402397, -0.3941513470775634,
           -0.29918000715316884, -0.20119409399743451, -0.1011420669187175,
           0.0, 0.1011420669187175, 0.20119409399743451,
           0.29918000715316884, 0.3941513470775634, 0.4850818636402397,
           0.5709721726085388, 0.650996741297417, 0.7244177313601701,
           0.790418501442466, 0.8482065834104272, 0.8972645323440819,
           0.937273392400706, 0.9677390756791391, 0.9879925180204854,
           0.9980022986933971]

      w = [0.005377479872923349, 0.015007947329316122, 0.02546084732671532,
           0.03534636079137585, 0.04458975132476488, 0.05348152469092809,
           0.06200956780067064, 0.06985412131872826, 0.07684968075772038,
           0.08308050282313302, 0.08856444305621176, 0.09312659817082532,
           0.09664272698362368, 0.09917359872179196, 0.10076984552387559,
           0.10133000701479154, 0.10076984552387559, 0.09917359872179196,
           0.09664272698362368, 0.09312659817082532, 0.08856444305621176,
           0.08308050282313302, 0.07684968075772038, 0.06985412131872826,
           0.06200956780067064, 0.05348152469092809, 0.04458975132476488,
           0.03534636079137585, 0.02546084732671532, 0.015007947329316122,
           0.005377479872923349]

    when 41
      # g20k41

      z = [-0.9988590315882777, -0.9931285991850949, -0.9815078774502503,
           -0.9639719272779138, -0.9408226338317548, -0.912234428251326,
           -0.878276811252282, -0.8391169718222188, -0.7950414288375512,
           -0.7463319064601508, -0.6932376563347514, -0.636053680726515,
           -0.5751404468197103, -0.5108670019508271, -0.4435931752387251,
           -0.37370608871541955, -0.301627868114913, -0.22778585114164507,
           -0.15260546524092267, -0.07652652113349734, 0.0,
           0.07652652113349734, 0.15260546524092267, 0.22778585114164507,
           0.301627868114913, 0.37370608871541955, 0.4435931752387251,
           0.5108670019508271, 0.5751404468197103, 0.636053680726515,
           0.6932376563347514, 0.7463319064601508, 0.7950414288375512,
           0.8391169718222188, 0.878276811252282, 0.912234428251326,
           0.9408226338317548, 0.9639719272779138, 0.9815078774502503,
           0.9931285991850949, 0.9988590315882777]

      w = [0.0030735837185205317, 0.008600269855642943,
           0.014626169256971253, 0.020388373461266523, 0.02588213360495116,
           0.0312873067770328, 0.036600169758200796, 0.041668873327973685,
           0.04643482186749767, 0.05094457392372869, 0.05519510534828599,
           0.05911140088063957, 0.06265323755478117, 0.06583459713361842,
           0.06864867292852161, 0.07105442355344407, 0.07303069033278667,
           0.07458287540049918, 0.07570449768455667, 0.07637786767208074,
           0.07660071191799965, 0.07637786767208074, 0.07570449768455667,
           0.07458287540049918, 0.07303069033278667, 0.07105442355344407,
           0.06864867292852161, 0.06583459713361842, 0.06265323755478117,
           0.05911140088063957, 0.05519510534828599, 0.05094457392372869,
           0.04643482186749767, 0.041668873327973685, 0.036600169758200796,
           0.0312873067770328, 0.02588213360495116, 0.020388373461266523,
           0.014626169256971253, 0.008600269855642943,
           0.0030735837185205317]

    when 61
      # g30k61

      z = [-0.9994844100504906, -0.9968934840746495, -0.9916309968704046,
           -0.9836681232797472, -0.9731163225011262, -0.9600218649683075,
           -0.94437444474856, -0.9262000474292743, -0.9055733076999078,
           -0.8825605357920527, -0.8572052335460612, -0.8295657623827684,
           -0.799727835821839, -0.7677774321048262, -0.7337900624532268,
           -0.6978504947933158, -0.6600610641266269, -0.6205261829892429,
           -0.5793452358263617, -0.5366241481420199, -0.49248046786177857,
           -0.44703376953808915, -0.4004012548303944, -0.3527047255308781,
           -0.30407320227362505, -0.25463692616788985,
           -0.20452511668230988, -0.15386991360858354,
           -0.10280693796673702, -0.0514718425553177, 0.0,
           0.0514718425553177, 0.10280693796673702, 0.15386991360858354,
           0.20452511668230988, 0.25463692616788985, 0.30407320227362505,
           0.3527047255308781, 0.4004012548303944, 0.44703376953808915,
           0.49248046786177857, 0.5366241481420199, 0.5793452358263617,
           0.6205261829892429, 0.6600610641266269, 0.6978504947933158,
           0.7337900624532268, 0.7677774321048262, 0.799727835821839,
           0.8295657623827684, 0.8572052335460612, 0.8825605357920527,
           0.9055733076999078, 0.9262000474292743, 0.94437444474856,
           0.9600218649683075, 0.9731163225011262, 0.9836681232797472,
           0.9916309968704046, 0.9968934840746495, 0.9994844100504906]

      w = [0.0013890136986770077, 0.003890461127099884,
           0.0066307039159312926, 0.009273279659517764,
           0.011823015253496341, 0.014369729507045804, 0.01692088918905327,
           0.019414141193942382, 0.021828035821609193, 0.0241911620780806,
           0.0265099548823331, 0.02875404876504129, 0.030907257562387762,
           0.03298144705748372, 0.034979338028060025, 0.03688236465182123,
           0.038678945624727595, 0.040374538951535956,
           0.041969810215164244, 0.04345253970135607, 0.04481480013316266,
           0.04605923827100699, 0.04718554656929915, 0.04818586175708713,
           0.04905543455502978, 0.04979568342707421, 0.05040592140278235,
           0.05088179589874961, 0.051221547849258774, 0.05142612853745902,
           0.05149472942945157, 0.05142612853745902, 0.051221547849258774,
           0.05088179589874961, 0.05040592140278235, 0.04979568342707421,
           0.04905543455502978, 0.04818586175708713, 0.04718554656929915,
           0.04605923827100699, 0.04481480013316266, 0.04345253970135607,
           0.041969810215164244, 0.040374538951535956,
           0.038678945624727595, 0.03688236465182123, 0.034979338028060025,
           0.03298144705748372, 0.030907257562387762, 0.02875404876504129,
           0.0265099548823331, 0.0241911620780806, 0.021828035821609193,
           0.019414141193942382, 0.01692088918905327, 0.014369729507045804,
           0.011823015253496341, 0.009273279659517764,
           0.0066307039159312926, 0.003890461127099884,
           0.0013890136986770077]

    else # using 15 point quadrature

      n = 15

      z = [-0.9914553711208126, -0.9491079123427585, -0.8648644233597691,
           -0.7415311855993945, -0.5860872354676911, -0.4058451513773972,
           -0.20778495500789848, 0.0, 0.20778495500789848,
           0.4058451513773972, 0.5860872354676911, 0.7415311855993945,
           0.8648644233597691, 0.9491079123427585, 0.9914553711208126]

      w = [0.022935322010529224, 0.06309209262997856, 0.10479001032225019,
           0.14065325971552592, 0.1690047266392679, 0.19035057806478542,
           0.20443294007529889, 0.20948214108472782, 0.20443294007529889,
           0.19035057806478542, 0.1690047266392679, 0.14065325971552592,
           0.10479001032225019, 0.06309209262997856, 0.022935322010529224]

  end

  sum = 0
  (0...n).each do |i|
    t = ((t1.to_f + t2) / 2.0) + (((t2 - t1) / 2.0) * z[i])
    sum += w[i] * yield(t)
  end

  ((t2 - t1) / 2.0) * sum
end

.has_gsl?Boolean

Check if GSL is available. Require the library if it is present. Return a boolean indicating its availability.

Returns:

  • (Boolean)

    Whether GSL is available.



186
187
188
189
190
191
192
193
194
195
196
197
198
# File 'lib/integration.rb', line 186

def has_gsl?
  gsl_available = '@@gsl'
  if class_variable_defined? gsl_available
    class_variable_get(gsl_available)
  else
    begin
      require 'gsl'
      class_variable_set(gsl_available, true)
    rescue LoadError
      class_variable_set(gsl_available, false)
    end
  end
end

.infinite?(value) ⇒ Boolean

Check if value is plus or minus infinity.

Parameters:

  • value

    Value to be tested.

Returns:

  • (Boolean)


52
53
54
# File 'lib/integration.rb', line 52

def infinite?(value)
  value == Integration::Infinity || value == Integration::MInfinity
end

.integrate(t1, t2, options = {}, &f) ⇒ Object

Get the integral for a function +f+, with bounds +t1+ and +t2+ given a hash of +options+. If Ruby/GSL is available, you can use +Integration::Minfinity+ and +Integration::Infinity+ as bounds. Method

Options are: [:tolerance] Maximum difference between real and calculated integral. Default: 1e-10. [:initial_step] Initial number of subdivisions. [:step] Subdivition increment on each iteration. [:method] Integration method.

Available methods are:

[:rectangle] for [:initial_step+:step*iteration] quadrilateral subdivisions. [:trapezoid] for [:initial_step+:step*iteration] trapezoid-al subdivisions. [:simpson] for [:initial_step+:step*iteration] parabolic subdivisions. [:adaptive_quadrature] for recursive appoximations until error [tolerance]. [:gauss] [:initial_step+:step*iteration] weighted subdivisons using translated -1 -> +1 endpoints. [:romberg] extrapolation of recursion approximation until error < [tolerance]. [:monte_carlo] make [:initial_step+:step*iteration] random samples, and check for above/below curve. [:qng] GSL QNG non-adaptive Gauss-Kronrod integration. [:qag] GSL QAG adaptive integration, with support for infinite bounds.



80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
# File 'lib/integration.rb', line 80

def integrate(t1, t2, options = {}, &f)
  inf_bounds = (infinite?(t1) || infinite?(t2))

  fail 'No function passed' unless block_given?
  fail 'Non-numeric bounds' unless ((t1.is_a? Numeric) && (t2.is_a? Numeric)) || inf_bounds

  if inf_bounds
    lower_bound = t1
    upper_bound = t2
    options[:method] = :qag if options[:method].nil?
  else
    lower_bound = [t1, t2].min
    upper_bound = [t1, t2].max
  end

  def_method = (Integration.has_gsl?) ? :qag : :simpson
  default_opts = { tolerance: 1e-10, initial_step: 16, step: 16, method: def_method }
  options = default_opts.merge(options)

  if RUBY_METHODS.include? options[:method]
    fail "Ruby methods doesn't support infinity bounds" if inf_bounds
    integrate_ruby(lower_bound, upper_bound, options, &f)
  elsif GSL_METHODS.include? options[:method]
    integrate_gsl(lower_bound, upper_bound, options, &f)
  else
    fail "Unknown integration method \"#{options[:method]}\""
  end
end

.integrate_gsl(lower_bound, upper_bound, options, &f) ⇒ Object

Integrate using the GSL bindings.



110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
# File 'lib/integration.rb', line 110

def integrate_gsl(lower_bound, upper_bound, options, &f)
  f = GSL::Function.alloc(&f)
  method = options[:method]
  tolerance = options[:tolerance]

  if (method == :qag)
    w = GSL::Integration::Workspace.alloc

    val = if infinite?(lower_bound) && infinite?(upper_bound)
      f.qagi([tolerance, 0.0], 1000, w)
    elsif infinite?(lower_bound)
      f.qagil(upper_bound, [tolerance, 0], w)
    elsif infinite?(upper_bound)
      f.qagiu(lower_bound, [tolerance, 0], w)
    else
      f.qag([lower_bound, upper_bound], [tolerance, 0.0], GSL::Integration::GAUSS61, w)
    end

  elsif (method == :qng)
    val = f.qng([lower_bound, upper_bound], [tolerance, 0.0])

  else
    fail "Unknown integration method \"#{method}\""
  end

  val[0]
end

.integrate_ruby(lower_bound, upper_bound, options, &f) ⇒ Object



138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
# File 'lib/integration.rb', line 138

def integrate_ruby(lower_bound, upper_bound, options, &f)
  method = options[:method]
  tolerance = options[:tolerance]
  initial_step = options[:initial_step]
  step = options[:step]
  points = options[:points]

  begin
    method_obj = Integration.method(method.to_s.downcase)
  rescue
    raise "Unknown integration method \"#{method}\""
  end

  current_step = initial_step

  if [:adaptive_quadrature, :romberg, :gauss, :gauss_kronrod].include? method
    if (method == :gauss)
      initial_step = 10 if initial_step > 10
      tolerance = initial_step
      method_obj.call(lower_bound, upper_bound, tolerance, &f)
    elsif (method == :gauss_kronrod)
      initial_step = 10 if initial_step > 10
      tolerance = initial_step
      points = points unless points.nil?
      method_obj.call(lower_bound, upper_bound, tolerance, points, &f)
    else
      method_obj.call(lower_bound, upper_bound, tolerance, &f)
    end
  else
    value = method_obj.call(lower_bound, upper_bound, current_step, &f)
    previous = value + (tolerance * 2)
    diffs = []

    while (previous - value).abs > tolerance
      diffs.push((previous - value).abs)
      current_step += step
      previous = value
      value = method_obj.call(lower_bound, upper_bound, current_step, &f)
    end

    value
  end
end

.milne(t1, t2, n, &f) ⇒ Object

Milne's Method +n+ implies number of subdivisions Source: Abramowitz, M. and Stegun, I. A. (Eds.). Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing. New York: Dover, pp. 896-897, 1972.



85
86
87
88
89
90
91
92
# File 'lib/integration/methods.rb', line 85

def milne(t1, t2, n, &f)
  d = (t2 - t1) / n.to_f
  ac = 0
  (0..n - 1).each do |i|
    ac += (d / 3.0) * (2 * f[t1 + i * d + d / 4] - f[t1 + i * d + d / 2] + 2 * f[t1 + i * d + 3 * d / 4])
  end
  ac
end

.monte_carlo(t1, t2, n) ⇒ Object

Monte Carlo

Uses a non-deterministic approach for calculation of definite integrals. Estimates the integral by randomly choosing points in a set and then calculating the number of points that fall in the desired area.



376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
# File 'lib/integration/methods.rb', line 376

def monte_carlo(t1, t2, n)
  width = (t2 - t1).to_f
  height = nil
  vals = []
  n.times do
    t = t1 + (rand * width)
    ft = yield(t)
    height = ft if height.nil? || ft > height
    vals << ft
  end
  area_ratio = 0
  vals.each do |ft|
    area_ratio += (ft / height.to_f) / n.to_f
  end
  (width * height) * area_ratio
end

.open_trapezoid(t1, t2, n, &f) ⇒ Object

Open Trapezoid method +n+ implies number of subdivisions Values computed at mid point and end point instead of starting points



70
71
72
73
74
75
76
77
# File 'lib/integration/methods.rb', line 70

def open_trapezoid(t1, t2, n, &f)
  d = (t2 - t1) / n.to_f
  ac = 0
  (0..n - 1).each do |i|
    ac += (d / 2.0) * (f[t1 + i * d + d / 3] + f[t1 + i * d + 2 * d / 3])
  end
  ac
end

.rectangle(t1, t2, n, &f) ⇒ Object Also known as: midpoint

Rectangle method +n+ implies number of subdivisions Source:

  • Ayres : Outline of calculus


8
9
10
11
12
13
# File 'lib/integration/methods.rb', line 8

def rectangle(t1, t2, n, &f)
  d = (t2 - t1) / n.to_f
  n.times.inject(0) do|ac, i|
    ac + f[t1 + d * (i + 0.5)]
  end * d
end

.romberg(a, b, tolerance, max_iter = 20) ⇒ Object

Romberg Method: It is obtained by applying extrapolation techniques repeatedly on the trapezoidal rule



351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
# File 'lib/integration/methods.rb', line 351

def romberg(a, b, tolerance, max_iter = 20)
  # NOTE one-based arrays are used for convenience
  h = b.to_f - a
  close = 1
  r = [[(h / 2) * (yield(a) + yield(b))]]
  j = 0
  hn = ->(n) { h / (2**n) }
  while j <= max_iter && tolerance < close
    j += 1
    r.push((j + 1).times.map { [] })
    ul = 2**(j - 1)
    r[j][0] = r[j - 1][0] / 2.0 + hn[j] * (1..ul).inject(0) { |ac, k| ac + yield(a + (2 * k - 1) * hn[j]) }
    (1..j).each do |k|
      r[j][k] = ((4**k) * r[j][k - 1] - r[j - 1][k - 1]) / ((4**k) - 1)
    end
    close = (r[j][j] - r[j - 1][j - 1])
  end
  r[j][j]
end

.simpson(t1, t2, n, &f) ⇒ Object

Simpson's rule +n+ implies number of subdivisions Source:

  • Ayres : Outline of calculus


32
33
34
35
36
37
38
39
# File 'lib/integration/methods.rb', line 32

def simpson(t1, t2, n, &f)
  n += 1 unless n.even?
  d = (t2 - t1) / n.to_f
  out = (d / 3.0) * (f[t1.to_f].to_f + ((1..(n - 1)).inject(0) do|ac, i|
    ac + ((i.even?) ? 2 : 4) * f[t1 + d * i]
  end) + f[t2.to_f].to_f)
  out
end

.simpson3by8(t1, t2, n, &f) ⇒ Object

Simpson's 3/8 Rule +n+ implies number of subdivisions Source:

  • Burden, Richard L. and Faires, J. Douglas (2000): Numerical Analysis (7th ed.). Brooks/Cole


45
46
47
48
49
50
51
52
# File 'lib/integration/methods.rb', line 45

def simpson3by8(t1, t2, n, &f)
  d = (t2 - t1) / n.to_f
  ac = 0
  (0..n - 1).each do |i|
    ac += (d / 8.0) * (f[t1 + i * d] + 3 * f[t1 + i * d + d / 3] + 3 * f[t1 + i * d + 2 * d / 3] + f[t1 + (i + 1) * d])
  end
  ac
end

.trapezoid(t1, t2, n, &f) ⇒ Object

Trapezoid method +n+ implies number of subdivisions Source:

  • Ayres : Outline of calculus


21
22
23
24
25
26
# File 'lib/integration/methods.rb', line 21

def trapezoid(t1, t2, n, &f)
  d = (t2 - t1) / n.to_f
  (d / 2.0) * (f[t1] + 2 * (1..(n - 1)).inject(0) do|ac, i|
    ac + f[t1 + d * i]
  end + f[t2])
end