python-openopt 0.38+svn1589-1 / usr / share / pyshared / openopt / tests / chain.py

This file is indexed.

/usr/share/pyshared/openopt/tests/chain.py is in python-openopt 0.38+svn1589-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below. 

  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
"""
The test is related to obsolete OpenOpt version and doesn't work for now
It was used to write an article related to numerical optimization.
"""

from numpy import *
from openopt import *
#from pylab import *

TestCollection = xrange(5, 26)
#TestCollection = xrange(5, 11)
#TestCollection = [5]
TestCollection = [21]

P = 0
#P = 5e0
xyCoordsAlwaysExist = False

solvers = ['ralg', 'ipopt','scipy_cobyla']
solvers = ['ralg', 'ipopt']
solvers = ['ralg']
#solvers = ['ipopt']
PLOT = 0
Results = {}

contol = 1e-8
xl, yl = 0, 15
yThreshold = -1.0


LeftPointForceX  = 10.00
LeftPointForceY = -3000.00

#maxY = array([yl, 10.0912237188, 5.81736169843, 2.38892359538,-0.528509686607, -3.61027100044,-7.21539449067,-11.0833587513,-14.5626399121,-17.2632362225,-19.3451872935,-21.3081000933,-23.4618325667,-25.641568098,-27.3915251977,-28.4811563214,-29.0663497915,-29.387570995,-29.4874690862,-29.2098877982,-28.3834712745,-27.140937007,-25.7876146522,-24.3568516911,-22.5652835804,-20.1292934475]) + yThreshold

maxY = array([inf, 10.0399786566, 5.61033453009, 2.06002552258, -0.96112448303, -4.13674736433, -7.84375429882, -11.8094052553, -14.9106560481, -17.1598341467, -18.7959512268, -20.240402618, -21.6919090157, -22.9270315304, -23.5750132311, -23.5920241002, -23.2285382362, -22.6103728598, -21.5974976575, -19.9739484665, -17.7441144032, -15.2828386661, -12.9735192373, -10.7235600474, -8.05824588682, -4.62542867851]) + yThreshold
#maxY[1] -= 100

for n in TestCollection:

    oovarInit = oovar('leftPointForces', v0 = [LeftPointForceX, LeftPointForceY], lb=[0, -inf])

    MaxForces = 100*sin(arange(n)) + 5000*ones(n)\
    + array([ -1.37163831e+03,  -1.60694848e+03,  -1.74685759e+03,\
            -1.77099023e+03,  -1.76295446e+03,  -1.83232862e+03,\
            -2.01327466e+03,  -2.23292838e+03,  -2.37990432e+03,\
            -2.40786079e+03,  -2.37604148e+03,  -2.39220174e+03,\
            -2.51227773e+03,  -2.68825834e+03,  -2.81448159e+03,\
            -2.82791948e+03,  -2.76385327e+03,  -2.71873406e+03,\
            -2.75969731e+03,  -2.86237323e+03,  -2.93613565e+03,\
            -2.91260196e+03,  -2.80726185e+03,  -2.69990883e+03,\
            -2.65982371e+03,  -2.68285823e+03,  -2.69793719e+03,\
            -2.63722331e+03,  -2.49886567e+03,  -2.34512448e+03,\
            -2.24405190e+03,  -2.20790597e+03,  -2.18342816e+03,\
            -2.10368051e+03,  -1.95071662e+03,  -1.77036769e+03,\
            -1.62911926e+03,  -1.55345727e+03,  -1.50621766e+03,\
            -1.42212572e+03,  -1.26885753e+03,  -1.07614103e+03,\
            -9.07638583e+02,  -8.02510393e+02,  -7.39726398e+02,\
            -6.58001243e+02,  -5.12877756e+02,  -3.17504802e+02,\
            -1.30014677e+02,   9.99898976e-03])[:n]

    lengths = 5*ones(n)+cos(arange(n))#array([4, 3, 1, 2])
    masses = 15*ones(n)+4*cos(arange(n))#array([10, 15, 20])

    g = 10
    Fm = masses * g

    ########################################################
    s = [20, 20]
    AdditionalMasses = oovar('AdditionalMasses', v0=s + [(100.0-sum(s))/(n-len(s))]*(n-len(s)), lb=zeros(n))
    ########################################################

    ########################################################
    from blockMisc import *

    #def blockEngineFunc(inp, AdditionalMasses, y_limit, blockID):
    def blockEngineFunc(inp, AdditionalMasses, blockID):
        if blockID == 0:
            lFx, lFy, lx, ly, prevBlockForceThreshold, prev_yLimit = inp[0], inp[1], xl, yl, 0, 0
        else:
            lFx, lFy, lx, ly, prevBlockForceThreshold, prev_yLimit = inp[0], inp[1], inp[2], inp[3], inp[4], inp[5]

        prevBlockBroken = blockID>=0 and isnan(prevBlockForceThreshold)\
        or \
        (prevBlockForceThreshold > 0 and P == 0)\
        or \
        (prev_yLimit > 0 or isnan(prev_yLimit) and P == 0 and not xyCoordsAlwaysExist)

        # calculate output
        Fwhole = sqrt(lFx**2+lFy**2)
        ForceThreshold = (Fwhole - MaxForces[blockID]) / 1e4

        CurrentAdditionalMass = AdditionalMasses[blockID]
        rFy = lFy +Fm[blockID] + CurrentAdditionalMass*g # TODO : store Fm[i] inside blocks
        rFx = lFx# Fx are same along whole chain

        dx, dy = lengths[blockID] * lFx/ Fwhole, lengths[blockID] * lFy/ Fwhole
        rx = lx + dx
        ry = ly + dy

        yLimit = ly - maxY[blockID]

        if P != 0:
            projection, distance = project2ball(x = [lFx, lFy], radius=MaxForces[blockID], center = 0)
            ForceThreshold += P * distance / 1e4

            projection, distance = project2box(ly, -inf, maxY[blockID])
            if distance > 0:
                yLimit = P/1e4 * distance

        if prevBlockBroken or (prev_yLimit>0 and P == 0): #and not xyCoordsAlwaysExist:
            rx, ry, ForceThreshold, rFx, rFy, yLimit = nan, nan, nan, nan, nan, nan


        r = array((rFx, rFy, rx, ry, ForceThreshold, yLimit))

        return r

    #def derivative_blockEngineFunc(inp, AdditionalMasses, y_limit, blockID):
    def derivative_blockEngineFunc(inp, AdditionalMasses, blockID):
        # TODO: return nans if prev block is broken
        if blockID == 0:
            lFx, lFy, lx, ly, prevBlockForceThreshold, prev_yLimit = inp[0], inp[1], xl, yl, 0, 0
        else:
            lFx, lFy, lx, ly, prevBlockForceThreshold, prev_yLimit = inp[0], inp[1], inp[2], inp[3], inp[4], inp[5]

        if blockID == 0:
            nVars = 2 + len(AdditionalMasses)
        else:
            nVars = 6 + len(AdditionalMasses)

        #r = zeros((len(inp)+1, nVars))
        r = zeros((6, nVars))

        # d_ rFx /
        r[0, 0] = 1 # d_lFx

        # d_rFy/
        r[1, 1] = 1 # d_lFy
        r[1, len(inp) + blockID] = g
        # TODO: Check it (below)
        #elif blockID < n-1:
            #r[3, 5 + blockID] = g

        # d_rx /
        if blockID!=0: r[2, 2] = 1 # dlx
        Fwhole = sqrt(lFx**2+lFy**2)
        #r[2, 0] = lengths[blockID] / Fwhole # dlFx
        r[2, 0] = lengths[blockID] * lFy**2 / (Fwhole ** 3) # d_lFx
        r[2, 1] = - lengths[blockID] * lFx * lFy / (Fwhole ** 3) # d_lFy

        # d_ry /
        if blockID!=0: r[3, 3] = 1 # dly
        r[3, 0] = - lengths[blockID] * lFx * lFy / (Fwhole ** 3) # d_lFx, and is same to r[2,1]
        r[3, 1] = lengths[blockID] * lFx**2 / (Fwhole ** 3) # d_lFy

        # dForceThreshold
        r[4, 0] = lFx / Fwhole / 1e4 # / dlFx
        r[4, 1] = lFy / Fwhole / 1e4 # / dlFy

        # dYlimit
        if blockID != 0: r[5, 3] = 1.0 #d_Ylimit / d_yl

        if P != 0:
            projection, distance = project2ball(x = [lFx, lFy], radius=MaxForces[blockID], center = 0)
            if distance != 0:
                penalty_derivative = P/1e4 * project2ball_derivative(x = [lFx, lFy], radius=MaxForces[blockID], center = 0)
                r[4, 0] += penalty_derivative[0]
                r[4, 1] += penalty_derivative[1]

            projection, distance = project2box(ly, -inf, maxY[blockID])
            if distance > 0:
                r[5, 3] = P/1e4 * project2box_derivative(ly, -inf, maxY[blockID])

            #projection, distance = project2box(ly, maxY[blockID], inf)
            #yLimit = P * distance
        prevBlockBroken = blockID>=0 and isnan(prevBlockForceThreshold)\
        or \
        (prevBlockForceThreshold > 0 and P == 0)\
        or \
        (prev_yLimit > 0 or isnan(prev_yLimit) and P == 0 and not xyCoordsAlwaysExist)
        if prevBlockBroken:
            r *= nan
        return r

    ooFuncs, c = [], []
    constrYmax = []

    for i in xrange(n):
        oof = oofun(blockEngineFunc, args = copy(i), name = 'blockEngine'+str(i))

        if i == 0:
            oof.input = (oovarInit, AdditionalMasses)
        else:
            oof.input = (ooFuncs[i-1], AdditionalMasses)
        oof.d = derivative_blockEngineFunc
        ooFuncs.append(oof)
        # TODO: replace "4" by named output "ForceThreshold"

        c.append(oolin([0, 0, 0, 0, 1, 0], input = ooFuncs[copy(i)], name='maxForce'+str(i)))
        c.append(oolin([0, 0, 0, 0, 0, 1], input = ooFuncs[copy(i)], name='Ylimit'+str(i)))
        #c.append(y_limit)


#    y_limit = oofun(lambda *inputs: [inp[1]-maxY[i] for i, inp in enumerate(inputs)], input = ooFuncs, name='maxY')
#    def d_y_limit(*inputs):
#        r = zeros((n, len(inputs[0])*n))
#        for i in xrange(n):
#            r[i, len(inputs[0]) * i + 1] = 1
#        return r
#    y_limit.d = d_y_limit
#
#    c.append(y_limit)

    f = oofun(lambda z: z[0]**1.5, input = ooFuncs[-1], d = lambda z:[0, 0, 1.5*z[0]**0.5, 0, 0, 0], name = 'objFunc')
    #f = oofun(lambda z: z[0], input = ooFuncs[-1], d = lambda z:[0, 0, 1, 0, 0,0], name = 'objFunc')
    #f = oolin(array([0, 0, 1, 0, 0, 0]), input = ooFuncs[-1], name = 'objFunc')

    sumOfMasses = oofun(lambda z: 1-z.sum()/100.0, input=AdditionalMasses, d = lambda z: -ones(n)/100.0, name='sOm')
    c.append(sumOfMasses)

    colors = ['b', 'r', 'g', 'y', 'm', 'c']

    for j, solver in enumerate(solvers):

        p = NLP(f, c=c, goal = 'max', gtol = 1e-6,  plot=0, contol = contol, maxFunEvals = 1e10)
        def callback(p):
            print p.c(p.xk)
            return 0
        #p.callback = callback
        if solver == 'scipy_cobyla':
            p.f_iter = max((int(n/2), 5))

        if solver == 'ipopt':
            p.maxIter = 1500 - 40*n
        else:
            p.maxIter = 15000

        r = p.solve(solver,  plot=0, showFeas=1, maxTime = 150, iprint = -1, ftol=1e-6, xtol=1e-6)

        Results[(n, p.solver.__name__)] = r
        if r.isFeasible: msgF = '+'
        else: msgF = '-'
        print 'n=%d' % n, ('f=%3.2f'% r.ff)+'['+msgF+']','Time=%3.1f' % r.elapsed['solver_time']


        if PLOT:
            hold(1)
            for i, oof in enumerate(ooFuncs):
                if i == 0:
                    plot([xl, ooFuncs[0](p.xk)[0]], [yl, ooFuncs[0](p.xk)[1]], colors[j])
                    plot([xl, ooFuncs[0](p.x0)[0]], [yl, ooFuncs[0](p.x0)[1]], 'k')
                else:
                    plot([ooFuncs[i-1](p.xk)[0], ooFuncs[i](p.xk)[0]], [ooFuncs[i-1](p.xk)[1], ooFuncs[i](p.xk)[1]], colors[j])
                    plot([ooFuncs[i-1](p.x0)[0], ooFuncs[i](p.x0)[0]], [ooFuncs[i-1](p.x0)[1], ooFuncs[i](p.x0)[1]], 'k')
    if PLOT: show()

APT Browse - Built by Thomas Orozco.