github.com/jingcheng-WU/gonum@v0.9.1-0.20210323123734-f1a2a11a8f7b/optimize/functions/vlse.go (about) 1 // Copyright ©2017 The Gonum Authors. All rights reserved. 2 // Use of this source code is governed by a BSD-style 3 // license that can be found in the LICENSE file. 4 5 package functions 6 7 import "math" 8 9 // This file implements functions from the Virtual Library of Simulation Experiments. 10 // https://www.sfu.ca/~ssurjano/optimization.html 11 // In many cases gradients and Hessians have been added. In some cases, these 12 // are not defined at certain points or manifolds. The gradient in these locations 13 // has been set to 0. 14 15 // Ackley implements the Ackley function, a function of arbitrary dimension that 16 // has many local minima. It has a single global minimum of 0 at 0. Its typical 17 // domain is the hypercube of [-32.768, 32.768]^d. 18 // f(x) = -20 * exp(-0.2 sqrt(1/d sum_i x_i^2)) - exp(1/d sum_i cos(2π x_i)) + 20 + exp(1) 19 // where d is the input dimension. 20 // 21 // Reference: 22 // https://www.sfu.ca/~ssurjano/ackley.html (obtained June 2017) 23 type Ackley struct{} 24 25 func (Ackley) Func(x []float64) float64 { 26 var ss, sc float64 27 for _, v := range x { 28 ss += v * v 29 sc += math.Cos(2 * math.Pi * v) 30 } 31 id := 1 / float64(len(x)) 32 return -20*math.Exp(-0.2*math.Sqrt(id*ss)) - math.Exp(id*sc) + 20 + math.E 33 } 34 35 // Bukin6 implements Bukin's 6th function. The function is two-dimensional, with 36 // the typical domain as x_0 ∈ [-15, -5], x_1 ∈ [-3, 3]. The function has a unique 37 // global minimum at [-10, 1], and many local minima. 38 // f(x) = 100 * sqrt(|x_1 - 0.01*x_0^2|) + 0.01*|x_0+10| 39 // Reference: 40 // https://www.sfu.ca/~ssurjano/bukin6.html (obtained June 2017) 41 type Bukin6 struct{} 42 43 func (Bukin6) Func(x []float64) float64 { 44 if len(x) != 2 { 45 panic(badInputDim) 46 } 47 return 100*math.Sqrt(math.Abs(x[1]-0.01*x[0]*x[0])) + 0.01*math.Abs(x[0]+10) 48 } 49 50 // CamelThree implements the three-hump camel function, a two-dimensional function 51 // with three local minima, one of which is global. 52 // The function is given by 53 // f(x) = 2*x_0^2 - 1.05*x_0^4 + x_0^6/6 + x_0*x_1 + x_1^2 54 // with the global minimum at 55 // x^* = (0, 0) 56 // f(x^*) = 0 57 // The typical domain is x_i ∈ [-5, 5] for all i. 58 // Reference: 59 // https://www.sfu.ca/~ssurjano/camel3.html (obtained December 2017) 60 type CamelThree struct{} 61 62 func (c CamelThree) Func(x []float64) float64 { 63 if len(x) != 2 { 64 panic("camelthree: dimension must be 2") 65 } 66 x0 := x[0] 67 x1 := x[1] 68 x02 := x0 * x0 69 x04 := x02 * x02 70 return 2*x02 - 1.05*x04 + x04*x02/6 + x0*x1 + x1*x1 71 } 72 73 // CamelSix implements the six-hump camel function, a two-dimensional function. 74 // with six local minima, two of which are global. 75 // The function is given by 76 // f(x) = (4 - 2.1*x_0^2 + x_0^4/3)*x_0^2 + x_0*x_1 + (-4 + 4*x_1^2)*x_1^2 77 // with the global minima at 78 // x^* = (0.0898, -0.7126), (-0.0898, 0.7126) 79 // f(x^*) = -1.0316 80 // The typical domain is x_0 ∈ [-3, 3], x_1 ∈ [-2, 2]. 81 // Reference: 82 // https://www.sfu.ca/~ssurjano/camel6.html (obtained December 2017) 83 type CamelSix struct{} 84 85 func (c CamelSix) Func(x []float64) float64 { 86 if len(x) != 2 { 87 panic("camelsix: dimension must be 2") 88 } 89 x0 := x[0] 90 x1 := x[1] 91 x02 := x0 * x0 92 x12 := x1 * x1 93 return (4-2.1*x02+x02*x02/3)*x02 + x0*x1 + (-4+4*x12)*x12 94 } 95 96 // CrossInTray implements the cross-in-tray function. The cross-in-tray function 97 // is a two-dimensional function with many local minima, and four global minima 98 // at (±1.3491, ±1.3491). The function is typically evaluated in the square 99 // [-10,10]^2. 100 // f(x) = -0.001(|sin(x_0)sin(x_1)exp(|100-sqrt((x_0^2+x_1^2)/π)|)|+1)^0.1 101 // Reference: 102 // https://www.sfu.ca/~ssurjano/crossit.html (obtained June 2017) 103 type CrossInTray struct{} 104 105 func (CrossInTray) Func(x []float64) float64 { 106 if len(x) != 2 { 107 panic(badInputDim) 108 } 109 x0 := x[0] 110 x1 := x[1] 111 exp := math.Abs(100 - math.Sqrt((x0*x0+x1*x1)/math.Pi)) 112 return -0.0001 * math.Pow(math.Abs(math.Sin(x0)*math.Sin(x1)*math.Exp(exp))+1, 0.1) 113 } 114 115 // DixonPrice implements the DixonPrice function, a function of arbitrary dimension 116 // Its typical domain is the hypercube of [-10, 10]^d. 117 // The function is given by 118 // f(x) = (x_0-1)^2 + \sum_{i=1}^{d-1} (i+1) * (2*x_i^2-x_{i-1})^2 119 // where d is the input dimension. There is a single global minimum, which has 120 // a location and value of 121 // x_i^* = 2^{-(2^{i+1}-2)/(2^{i+1})} for i = 0, ..., d-1. 122 // f(x^*) = 0 123 // Reference: 124 // https://www.sfu.ca/~ssurjano/dixonpr.html (obtained June 2017) 125 type DixonPrice struct{} 126 127 func (DixonPrice) Func(x []float64) float64 { 128 xp := x[0] 129 v := (xp - 1) * (xp - 1) 130 for i := 1; i < len(x); i++ { 131 xn := x[i] 132 tmp := (2*xn*xn - xp) 133 v += float64(i+1) * tmp * tmp 134 xp = xn 135 } 136 return v 137 } 138 139 // DropWave implements the drop-wave function, a two-dimensional function with 140 // many local minima and one global minimum at 0. The function is typically evaluated 141 // in the square [-5.12, 5.12]^2. 142 // f(x) = - (1+cos(12*sqrt(x0^2+x1^2))) / (0.5*(x0^2+x1^2)+2) 143 // Reference: 144 // https://www.sfu.ca/~ssurjano/drop.html (obtained June 2017) 145 type DropWave struct{} 146 147 func (DropWave) Func(x []float64) float64 { 148 if len(x) != 2 { 149 panic(badInputDim) 150 } 151 x0 := x[0] 152 x1 := x[1] 153 num := 1 + math.Cos(12*math.Sqrt(x0*x0+x1*x1)) 154 den := 0.5*(x0*x0+x1*x1) + 2 155 return -num / den 156 } 157 158 // Eggholder implements the Eggholder function, a two-dimensional function with 159 // many local minima and one global minimum at [512, 404.2319]. The function 160 // is typically evaluated in the square [-512, 512]^2. 161 // f(x) = -(x_1+47)*sin(sqrt(|x_1+x_0/2+47|))-x_1*sin(sqrt(|x_0-(x_1+47)|)) 162 // Reference: 163 // https://www.sfu.ca/~ssurjano/egg.html (obtained June 2017) 164 type Eggholder struct{} 165 166 func (Eggholder) Func(x []float64) float64 { 167 if len(x) != 2 { 168 panic(badInputDim) 169 } 170 x0 := x[0] 171 x1 := x[1] 172 return -(x1+47)*math.Sin(math.Sqrt(math.Abs(x1+x0/2+47))) - 173 x0*math.Sin(math.Sqrt(math.Abs(x0-x1-47))) 174 } 175 176 // GramacyLee implements the Gramacy-Lee function, a one-dimensional function 177 // with many local minima. The function is typically evaluated on the domain [0.5, 2.5]. 178 // f(x) = sin(10πx)/(2x) + (x-1)^4 179 // Reference: 180 // https://www.sfu.ca/~ssurjano/grlee12.html (obtained June 2017) 181 type GramacyLee struct{} 182 183 func (GramacyLee) Func(x []float64) float64 { 184 if len(x) != 1 { 185 panic(badInputDim) 186 } 187 x0 := x[0] 188 return math.Sin(10*math.Pi*x0)/(2*x0) + math.Pow(x0-1, 4) 189 } 190 191 // Griewank implements the Griewank function, a function of arbitrary dimension that 192 // has many local minima. It has a single global minimum of 0 at 0. Its typical 193 // domain is the hypercube of [-600, 600]^d. 194 // f(x) = \sum_i x_i^2/4000 - \prod_i cos(x_i/sqrt(i)) + 1 195 // where d is the input dimension. 196 // 197 // Reference: 198 // https://www.sfu.ca/~ssurjano/griewank.html (obtained June 2017) 199 type Griewank struct{} 200 201 func (Griewank) Func(x []float64) float64 { 202 var ss float64 203 pc := 1.0 204 for i, v := range x { 205 ss += v * v 206 pc *= math.Cos(v / math.Sqrt(float64(i+1))) 207 } 208 return ss/4000 - pc + 1 209 } 210 211 // HolderTable implements the Holder table function. The Holder table function 212 // is a two-dimensional function with many local minima, and four global minima 213 // at (±8.05502, ±9.66459). The function is typically evaluated in the square [-10,10]^2. 214 // f(x) = -|sin(x_0)cos(x1)exp(|1-sqrt(x_0^2+x1^2)/π|)| 215 // Reference: 216 // https://www.sfu.ca/~ssurjano/holder.html (obtained June 2017) 217 type HolderTable struct{} 218 219 func (HolderTable) Func(x []float64) float64 { 220 if len(x) != 2 { 221 panic(badInputDim) 222 } 223 x0 := x[0] 224 x1 := x[1] 225 return -math.Abs(math.Sin(x0) * math.Cos(x1) * math.Exp(math.Abs(1-math.Sqrt(x0*x0+x1*x1)/math.Pi))) 226 } 227 228 // Langermann2 implements the two-dimensional version of the Langermann function. 229 // The Langermann function has many local minima. The function is typically 230 // evaluated in the square [0,10]^2. 231 // f(x) = \sum_1^5 c_i exp(-(1/π)\sum_{j=1}^2(x_j-A_{ij})^2) * cos(π\sum_{j=1}^2 (x_j - A_{ij})^2) 232 // c = [5]float64{1,2,5,2,3} 233 // A = [5][2]float64{{3,5},{5,2},{2,1},{1,4},{7,9}} 234 // Reference: 235 // https://www.sfu.ca/~ssurjano/langer.html (obtained June 2017) 236 type Langermann2 struct{} 237 238 func (Langermann2) Func(x []float64) float64 { 239 if len(x) != 2 { 240 panic(badInputDim) 241 } 242 var ( 243 c = [5]float64{1, 2, 5, 2, 3} 244 A = [5][2]float64{{3, 5}, {5, 2}, {2, 1}, {1, 4}, {7, 9}} 245 ) 246 var f float64 247 for i, cv := range c { 248 var ss float64 249 for j, av := range A[i] { 250 xja := x[j] - av 251 ss += xja * xja 252 } 253 f += cv * math.Exp(-(1/math.Pi)*ss) * math.Cos(math.Pi*ss) 254 } 255 return f 256 } 257 258 // Levy implements the Levy function, a function of arbitrary dimension that 259 // has many local minima. It has a single global minimum of 0 at 1. Its typical 260 // domain is the hypercube of [-10, 10]^d. 261 // f(x) = sin^2(π*w_0) + \sum_{i=0}^{d-2}(w_i-1)^2*[1+10sin^2(π*w_i+1)] + 262 // (w_{d-1}-1)^2*[1+sin^2(2π*w_{d-1})] 263 // w_i = 1 + (x_i-1)/4 264 // where d is the input dimension. 265 // 266 // Reference: 267 // https://www.sfu.ca/~ssurjano/levy.html (obtained June 2017) 268 type Levy struct{} 269 270 func (Levy) Func(x []float64) float64 { 271 w1 := 1 + (x[0]-1)/4 272 s1 := math.Sin(math.Pi * w1) 273 sum := s1 * s1 274 for i := 0; i < len(x)-1; i++ { 275 wi := 1 + (x[i]-1)/4 276 s := math.Sin(math.Pi*wi + 1) 277 sum += (wi - 1) * (wi - 1) * (1 + 10*s*s) 278 } 279 wd := 1 + (x[len(x)-1]-1)/4 280 sd := math.Sin(2 * math.Pi * wd) 281 return sum + (wd-1)*(wd-1)*(1+sd*sd) 282 } 283 284 // Levy13 implements the Levy-13 function, a two-dimensional function 285 // with many local minima. It has a single global minimum of 0 at 1. Its typical 286 // domain is the square [-10, 10]^2. 287 // f(x) = sin^2(3π*x_0) + (x_0-1)^2*[1+sin^2(3π*x_1)] + (x_1-1)^2*[1+sin^2(2π*x_1)] 288 // Reference: 289 // https://www.sfu.ca/~ssurjano/levy13.html (obtained June 2017) 290 type Levy13 struct{} 291 292 func (Levy13) Func(x []float64) float64 { 293 if len(x) != 2 { 294 panic(badInputDim) 295 } 296 x0 := x[0] 297 x1 := x[1] 298 s0 := math.Sin(3 * math.Pi * x0) 299 s1 := math.Sin(3 * math.Pi * x1) 300 s2 := math.Sin(2 * math.Pi * x1) 301 return s0*s0 + (x0-1)*(x0-1)*(1+s1*s1) + (x1-1)*(x1-1)*(1+s2*s2) 302 } 303 304 // Rastrigin implements the Rastrigen function, a function of arbitrary dimension 305 // that has many local minima. It has a single global minimum of 0 at 0. Its typical 306 // domain is the hypercube of [-5.12, 5.12]^d. 307 // f(x) = 10d + \sum_i [x_i^2 - 10cos(2π*x_i)] 308 // where d is the input dimension. 309 // 310 // Reference: 311 // https://www.sfu.ca/~ssurjano/rastr.html (obtained June 2017) 312 type Rastrigin struct{} 313 314 func (Rastrigin) Func(x []float64) float64 { 315 sum := 10 * float64(len(x)) 316 for _, v := range x { 317 sum += v*v - 10*math.Cos(2*math.Pi*v) 318 } 319 return sum 320 } 321 322 // Schaffer2 implements the second Schaffer function, a two-dimensional function 323 // with many local minima. It has a single global minimum of 0 at 0. Its typical 324 // domain is the square [-100, 100]^2. 325 // f(x) = 0.5 + (sin^2(x_0^2-x_1^2)-0.5) / (1+0.001*(x_0^2+x_1^2))^2 326 // Reference: 327 // https://www.sfu.ca/~ssurjano/schaffer2.html (obtained June 2017) 328 type Schaffer2 struct{} 329 330 func (Schaffer2) Func(x []float64) float64 { 331 if len(x) != 2 { 332 panic(badInputDim) 333 } 334 x0 := x[0] 335 x1 := x[1] 336 s := math.Sin(x0*x0 - x1*x1) 337 den := 1 + 0.001*(x0*x0+x1*x1) 338 return 0.5 + (s*s-0.5)/(den*den) 339 } 340 341 // Schaffer4 implements the fourth Schaffer function, a two-dimensional function 342 // with many local minima. Its typical domain is the square [-100, 100]^2. 343 // f(x) = 0.5 + (cos(sin(|x_0^2-x_1^2|))-0.5) / (1+0.001*(x_0^2+x_1^2))^2 344 // Reference: 345 // https://www.sfu.ca/~ssurjano/schaffer4.html (obtained June 2017) 346 type Schaffer4 struct{} 347 348 func (Schaffer4) Func(x []float64) float64 { 349 if len(x) != 2 { 350 panic(badInputDim) 351 } 352 x0 := x[0] 353 x1 := x[1] 354 den := 1 + 0.001*(x0*x0+x1*x1) 355 return 0.5 + (math.Cos(math.Sin(math.Abs(x0*x0-x1*x1)))-0.5)/(den*den) 356 } 357 358 // Schwefel implements the Schwefel function, a function of arbitrary dimension 359 // that has many local minima. Its typical domain is the hypercube of [-500, 500]^d. 360 // f(x) = 418.9829*d - \sum_i x_i*sin(sqrt(|x_i|)) 361 // where d is the input dimension. 362 // 363 // Reference: 364 // https://www.sfu.ca/~ssurjano/schwef.html (obtained June 2017) 365 type Schwefel struct{} 366 367 func (Schwefel) Func(x []float64) float64 { 368 var sum float64 369 for _, v := range x { 370 sum += v * math.Sin(math.Sqrt(math.Abs(v))) 371 } 372 return 418.9829*float64(len(x)) - sum 373 } 374 375 // Shubert implements the Shubert function, a two-dimensional function 376 // with many local minima and many global minima. Its typical domain is the 377 // square [-10, 10]^2. 378 // f(x) = (sum_{i=1}^5 i cos((i+1)*x_0+i)) * (\sum_{i=1}^5 i cos((i+1)*x_1+i)) 379 // Reference: 380 // https://www.sfu.ca/~ssurjano/shubert.html (obtained June 2017) 381 type Shubert struct{} 382 383 func (Shubert) Func(x []float64) float64 { 384 if len(x) != 2 { 385 panic(badInputDim) 386 } 387 x0 := x[0] 388 x1 := x[1] 389 var s0, s1 float64 390 for i := 1.0; i <= 5.0; i++ { 391 s0 += i * math.Cos((i+1)*x0+i) 392 s1 += i * math.Cos((i+1)*x1+i) 393 } 394 return s0 * s1 395 }