/usr/share/pyshared/openopt/solvers/nlopt/NLOPT_AUX.py is in python-openopt 0.34+svn1146-1build1.
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 | #from numpy import asarray, ones, all, isfinite, copy, nan, concatenate, array, dot
#from openopt.kernel.ooMisc import WholeRepr2LinConst, xBounds2Matrix
from openopt.kernel.setDefaultIterFuncs import SOLVED_WITH_UNIMPLEMENTED_OR_UNKNOWN_REASON, SMALL_DELTA_X, SMALL_DELTA_F
from numpy import isfinite, asscalar, asfarray, abs, copy, isinf, ones, array
import nlopt
def NLOPT_AUX(p, solver, opts=None):
def myfunc(x, grad):
#if p.istop != 0: raise nlopt.FORCED_STOP
if grad.size > 0:
grad[:]= p.df(x.copy())
return asscalar(p.f(x))
# TODO: add other local nlopt optimizers that cannot handle equality constraints
if solver in [nlopt.LD_MMA] and (p.nbeq != 0 or p.nh != 0):
opt = nlopt.opt(nlopt.LD_AUGLAG, p.n)
opt2 = nlopt.opt(solver, p.n)
setStopCriteria(opt2, p, reduce=10.0)
opt.set_local_optimizer(opt2)
elif solver == nlopt.G_MLSL_LDS:
opt = nlopt.opt(nlopt.G_MLSL_LDS, p.n)
opt2 = nlopt.opt(nlopt.LD_TNEWTON_PRECOND_RESTART, p.n)
setStopCriteria(opt2, p, reduce=10.0)
opt.set_local_optimizer(opt2)
else:
opt = nlopt.opt(solver, p.n)
if opts is not None:
for option, val in opts.items():
getattr(opt, option)(val)
opt.set_min_objective(myfunc)
if any(p.lb==p.ub): p.pWarn('nlopt solvers badly handle problems with variables fixed via setting lb=ub')
lb = [elem if isfinite(elem) else float(elem) for elem in p.lb.tolist()]
ub = [elem if isfinite(elem) else float(elem) for elem in p.ub.tolist()]
if any(isfinite(lb)): opt.set_lower_bounds(lb)
if any(isfinite(ub)): opt.set_upper_bounds(ub)
# TODO: A, Aeq
if p.nc > 0:
def c(result, x, grad):
if grad.size > 0:
grad[:] = p.dc(x.copy())
result[:] = p.c(x)
opt.add_inequality_mconstraint(c, array([p.contol]*p.nc).copy())
if p.nb > 0:
def c_lin(result, x, grad):
if grad.size > 0:
grad[:] = p.A.copy()
result[:] = p._get_AX_Less_B_residuals(x)
opt.add_inequality_mconstraint(c_lin, array([p.contol]*p.nb).copy())
if p.nh > 0:
def h(result, x, grad):
if grad.size > 0:
grad[:] = p.dh(x.copy())
result[:] = p.h(x).copy()
opt.add_equality_mconstraint(h, array([p.contol]*p.nh))
if p.nbeq > 0:
def h_lin(result, x, grad):
if grad.size > 0:
grad[:] = p.Aeq.copy()
result[:] = p._get_AeqX_eq_Beq_residuals(x)
opt.add_equality_mconstraint(h_lin, array([p.contol]*p.nbeq))
setStopCriteria(opt, p)
# others like fEnough, maxFunEvals, are handled by OO kernel
x0 = asfarray(p.x0).copy()
#lb2 = copy(lb)
#lb2[isinf(lb2)] = 0
LB = copy(lb) #+ 1e-15*(ones(p.n) + abs(lb2))
ind = x0 <= LB
x0[ind] = LB[ind]
#ub2 = copy(ub)
#ub2[isinf(ub2)] = 0
UB = copy(ub) #- 1e-15*(ones(p.n) + abs(ub2))
ind = x0 >= UB
x0[ind] = UB[ind]
#x0[ind] = p.lb[ind] + 1e-15*abs((max(1.0, p.lb[x0<p.lb])))
#x0[x0>p.ub] = p.ub[x0>p.ub] - 1e-15*min((1.0, abs(p.ub[x0>p.ub])))
# if p.solver.__name__ == 'auglag':
# opt.set_local_optimizer('mma')
try:
x = opt.optimize(x0.tolist()).copy()
#if p.point(x).betterThan(p.point(p.xk)):
#p.xk = x
except:
pass
p.xk = p._bestPoint.x
#x = p.xk
#p.iterfcn(x)
iStop = opt.get_stopval()
if p.istop == 0:
if iStop == nlopt.XTOL_REACHED:
p.istop, p.msg = SMALL_DELTA_X, '|| X[k] - X[k-1] || < xtol'
elif iStop == nlopt.FTOL_REACHED:
p.istop, p.msg = SMALL_DELTA_F, '|| F[k] - F[k-1] || < ftol'
else:
p.istop = SOLVED_WITH_UNIMPLEMENTED_OR_UNKNOWN_REASON
# p._opt = opt
def setStopCriteria(opt, p, reduce=1.0):
opt.set_xtol_abs(p.xtol/reduce)
opt.set_ftol_abs(p.ftol/reduce)
opt.set_maxeval(p.maxFunEvals)
#opt.set_maxiter(p.maxIter)
if isfinite(p.maxTime):
opt.set_maxtime(p.maxTime)
# opt.set_xtol_rel(1e-1)
# opt.set_ftol_rel(1e-1)
|