Skip to content
This repository
tree: 6fc6376668
Fetching contributors…

Cannot retrieve contributors at this time

file 336 lines (295 sloc) 11.163 kb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336
# Original Author: Travis Oliphant 2002
# Bug-fixes in 2006 by Tim Leslie
# Bug-fixes and changes by Carlos Lopez 2012
#

import numpy
from numpy import asarray, tan, exp, ones, squeeze, sign, \
     all, log, sqrt, pi, shape, array, minimum, where
from numpy import random

__all__ = ['anneal']

_double_min = numpy.finfo(float).min
_double_max = numpy.finfo(float).max
class base_schedule(object):
    def __init__(self):
        self.dwell = 20
        self.learn_rate = 0.5
        self.lower = -10
        self.upper = 10
        self.Ninit = 50
        self.accepted = 0
        self.tests = 0
        self.feval = 0
        self.k = 0
        self.T = None
        self.cvar = .05 # 5% coefficient of variation

    def init(self, **options):
        self.__dict__.update(options)
        self.lower = asarray(self.lower)
        self.lower = where(self.lower == numpy.NINF, -_double_max, self.lower)
        self.upper = asarray(self.upper)
        self.upper = where(self.upper == numpy.PINF, _double_max, self.upper)
        self.k = 0
        self.accepted = 0
        self.feval = 0
        self.tests = 0

    def getstart_temp(self, best_state):
        """ Find a matching starting temperature and starting parameters vector
i.e. find x0 such that func(x0) = T0.

Parameters
----------
best_state : _state
A _state object to store the function value and x0 found.

Returns
-------
x0 : array
The starting parameters vector.
"""
        print "============================================================"
        print "FINDING INITIAL TEMPERATURE WITH A COEFF OF VARIANCE OF", self.cvar
        print "============================================================"

        assert(not self.dims is None)
        lrange = self.lower
        urange = self.upper
        cvar = self.cvar
        fmax = _double_min
        fmin = _double_max
        x0 = best_state.x
        for _ in range(self.Ninit):
            print "sampling T step:", _
            samp = squeeze(random.uniform(0, 1, size=self.dims)) - 0.5
            samp = samp/0.5
            varx0 = x0 * samp * cvar # random number within the cvar range
            x0 = x0 + varx0
            fval = self.func(x0, *self.args)
            self.feval += 1
            if fval > fmax:
                fmax = fval
            if fval < fmin:
                fmin = fval
                best_state.cost = fval
                best_state.x = array(x0)

        self.T0 = (fmax-fmin)*1.5
        print "================================="
        print "SET INITIAL TEMPERATURE TO:", self.T0
        print "================================="
        return best_state.x

    def accept_test(self, dE):
        T = self.T
        self.tests += 1
        if dE < 0:
            self.accepted += 1
            return 1
        p = exp(-dE*1.0/self.boltzmann/T)
        if (p > random.uniform(0.0, 1.0)):
            self.accepted += 1
            return 1
        return 0

    def update_guess(self, x0):
        pass

    def update_temp(self, x0):
        pass


# A schedule due to Lester Ingber
class fast_sa(base_schedule):
    def init(self, **options):
        self.__dict__.update(options)
        if self.m is None:
            self.m = 1.0
        if self.n is None:
            self.n = 1.0
        self.c = self.m * exp(-self.n * self.quench)

    def update_guess(self, x0):
        x0 = asarray(x0)
        u = squeeze(random.uniform(0.0, 1.0, size=self.dims))
        T = self.T
        y = sign(u-0.5)*T*((1+1.0/T)**abs(2*u-1)-1.0)
        xc = y*(self.upper - self.lower)
        xnew = x0 + xc
        return xnew

    def update_temp(self):
        self.T = self.T0*exp(-self.c * self.k**(self.quench))
        self.k += 1
        return

class cauchy_sa(base_schedule):
    def update_guess(self, x0):
        x0 = asarray(x0)
        numbers = squeeze(random.uniform(-pi/2, pi/2, size=self.dims))
        xc = self.learn_rate * self.T * tan(numbers)
        xnew = x0 + xc
        return xnew

    def update_temp(self):
        self.T = self.T0/(1+self.k)
        self.k += 1
        return

class boltzmann_sa(base_schedule):
    def update_guess(self, x0):
        std = minimum(sqrt(self.T)*ones(self.dims), (self.upper-self.lower)/3.0/self.learn_rate)
        x0 = asarray(x0)
        xc = squeeze(random.normal(0, 1.0, size=self.dims))

        xnew = x0 + xc*std*self.learn_rate
        return xnew

    def update_temp(self):
        self.k += 1
        self.T = self.T0 / log(self.k+1.0)
        return

class _state(object):
    def __init__(self):
        self.x = None
        self.cost = None

# TODO:
# allow for general annealing temperature profile
# in that case use update given by alpha and omega and
# variation of all previous updates and temperature?

# Simulated annealing

def anneal(func, x0, args=(), schedule='fast', full_output=0,
           T0=None, Tf=1e-12, maxeval=None, maxaccept=None, maxiter=400,
           boltzmann=1.0, learn_rate=0.5, feps=1e-6, quench=1.0, m=1.0, n=1.0,
           lower=-100, upper=100, dwell=50, cvar=0.05):
    """Minimize a function using simulated annealing.

Schedule is a schedule class implementing the annealing schedule.
Available ones are 'fast', 'cauchy', 'boltzmann'

Parameters
----------
func : callable f(x, *args)
Function to be optimized.
x0 : ndarray
Initial guess.
args : tuple
Extra parameters to `func`.
schedule : base_schedule
Annealing schedule to use (a class).
full_output : bool
Whether to return optional outputs.
T0 : float
Initial Temperature (estimated as 1.2 times the largest
cost-function deviation over random points in the range).
Tf : float
Final goal temperature.
maxeval : int
Maximum function evaluations.
maxaccept : int
Maximum changes to accept.
maxiter : int
Maximum cooling iterations.
learn_rate : float
Scale constant for adjusting guesses.
boltzmann : float
Boltzmann constant in acceptance test
(increase for less stringent test at each temperature).
feps : float
Stopping relative error tolerance for the function value in
last four coolings.
quench, m, n : float
Parameters to alter fast_sa schedule.
lower, upper : float or ndarray
Lower and upper bounds on `x`.
dwell : int
The number of times to search the space at each temperature.

Returns
-------
xmin : ndarray
Point giving smallest value found.
retval : int
Flag indicating stopping condition::

0 : Cooled to global optimum
1 : Cooled to final temperature
2 : Maximum function evaluations
3 : Maximum cooling iterations reached
4 : Maximum accepted query locations reached

Jmin : float
Minimum value of function found.
T : float
Final temperature.
feval : int
Number of function evaluations.
iters : int
Number of cooling iterations.
accept : int
Number of tests accepted.

"""
    x0 = asarray(x0)
    lower = asarray(lower)
    upper = asarray(upper)

    schedule = eval(schedule+'_sa()')
    # initialize the schedule
    schedule.init(dims=shape(x0),func=func,args=args,boltzmann=boltzmann,T0=T0,
                  learn_rate=learn_rate, lower=lower, upper=upper,
                  m=m, n=n, quench=quench, dwell=dwell, cvar=cvar)

    current_state, last_state, best_state = _state(), _state(), _state()
    if T0 is None:
        best_state.x = x0
        x0 = schedule.getstart_temp(best_state)
    else:
        best_state.x = None
        best_state.cost = 300e8

    last_state.x = asarray(x0).copy()
    fval = func(x0,*args)
    schedule.feval += 1
    last_state.cost = fval
    if last_state.cost < best_state.cost:
        best_state.cost = fval
        best_state.x = asarray(x0).copy()
    schedule.T = schedule.T0
    fqueue = [100, 300, 500, 700]
    iters = 0
    while 1:
        for n in range(dwell):
            current_state.x = schedule.update_guess(last_state.x)
            current_state.cost = func(current_state.x,*args)
            schedule.feval += 1

            dE = current_state.cost - last_state.cost
            if schedule.accept_test(dE):
                last_state.x = current_state.x.copy()
                last_state.cost = current_state.cost
                if last_state.cost < best_state.cost:
                    best_state.x = last_state.x.copy()
                    best_state.cost = last_state.cost
        schedule.update_temp()
        iters += 1
        # Stopping conditions
        # 0) last saved values of f from each cooling step
        # are all very similar (effectively cooled)
        # 1) Tf is set and we are below it
        # 2) maxeval is set and we are past it
        # 3) maxiter is set and we are past it
        # 4) maxaccept is set and we are past it

        fqueue.append(squeeze(last_state.cost))
        fqueue.pop(0)
        af = asarray(fqueue)*1.0
        if all(abs((af-af[0])/af[0]) < feps):
            retval = 0
            if abs(af[-1]-best_state.cost) > feps*10:
                retval = 5
                print "Warning: Cooled to %f at %s but this is not" \
                      % (squeeze(last_state.cost), str(squeeze(last_state.x))) \
                      + " the smallest point found."
            break
        if (Tf is not None) and (schedule.T < Tf):
            retval = 1
            break
        if (maxeval is not None) and (schedule.feval > maxeval):
            retval = 2
            break
        if (iters > maxiter):
            print "Warning: Maximum number of iterations exceeded."
            retval = 3
            break
        if (maxaccept is not None) and (schedule.accepted > maxaccept):
            retval = 4
            break

    if full_output:
        return best_state.x, best_state.cost, schedule.T, \
               schedule.feval, iters, schedule.accepted, retval
    else:
        return best_state.x, retval



if __name__ == "__main__":
    from numpy import cos
    # minimum expected at ~-0.195
    func = lambda x: cos(14.5*x-0.3) + (x+0.2)*x
    print anneal(func,1.0,full_output=1,upper=3.0,lower=-3.0,feps=1e-4,maxiter=2000,schedule='cauchy')
    print anneal(func,1.0,full_output=1,upper=3.0,lower=-3.0,feps=1e-4,maxiter=2000,schedule='fast')
    print anneal(func,1.0,full_output=1,upper=3.0,lower=-3.0,feps=1e-4,maxiter=2000,schedule='boltzmann')

    # minimum expected at ~[-0.195, -0.1]
    func = lambda x: cos(14.5*x[0]-0.3) + (x[1]+0.2)*x[1] + (x[0]+0.2)*x[0]
    print anneal(func,[1.0, 1.0],full_output=1,upper=[3.0, 3.0],lower=[-3.0, -3.0],feps=1e-4,maxiter=2000,schedule='cauchy')
    print anneal(func,[1.0, 1.0],full_output=1,upper=[3.0, 3.0],lower=[-3.0, -3.0],feps=1e-4,maxiter=2000,schedule='fast')
    print anneal(func,[1.0, 1.0],full_output=1,upper=[3.0, 3.0],lower=[-3.0, -3.0],feps=1e-4,maxiter=2000,schedule='boltzmann')
Something went wrong with that request. Please try again.