def solve_dual(cap_array, fx, dfx, gm, Budg):
"""
Input:capacity array (e.g. vertices of U) or its nec.measure component, functions, gradients, global maximums of capacities
Finds the robust capacity, i.e. the solution of dual problem
Capacities assumed to be 2-MONOTONE
Output: v^r
see max_dual() for scipy version
"""
x0 = array([1. / shape(cap_array)[0] for i in range(shape(cap_array)[0])])
# print x0
Aeq = ones(shape(cap_array)[0])
beq = Budg
A = diag(-ones(shape(cap_array)[0]))
b = zeros(shape(cap_array)[0])
vmax = gm
# vmax = array([chq.Choquet(array(chq.max_choquet(v,Budg)),v) for v in cap_array])
def fprime(x, cap_array, Budg):
cap_int = dot(cap_array.T, x)
x_int = array(max_choquet(cap_int, fx, dfx, Budg))
ch_xr = array([Choquet(x_int, v, fx) for v in cap_array])
ch_diffs = ch_xr - vmax
return ch_diffs
def objfunc(x, cap_array, Budg):
cap_int = dot(cap_array.T, x)
x_int = array(max_choquet(cap_int, fx, dfx, Budg))
ch_xr = array([Choquet(x_int, v, fx) for v in cap_array])
ch_diffs = ch_xr - vmax
# print dot(x, ch_diffs)
return dot(x, ch_diffs)
objfunc_w = lambda x: objfunc(x, cap_array, Budg)
fprime_w = lambda x: fprime(x, cap_array, Budg)
p = NSP(objfunc_w,
x0,
df=fprime_w,
A=A,
b=b,
Aeq=Aeq,
beq=beq,
iprint=25,
maxIter=300,
maxFunEvals=1e7,
diffint=1e-10,
xtol=1e-9,
ftol=1e-7,
gtol=1e-5)
r = p.solve('ralg')
print r.xf, r.ff
cap_mix = dot(cap_array.T, array(r.xf))
return r.ff, cap_mix
def solve_dual(cap_array,fx,dfx,gm,Budg):
"""
Input:capacity array (e.g. vertices of U) or its nec.measure component, functions, gradients, global maximums of capacities
Finds the robust capacity, i.e. the solution of dual problem
Capacities assumed to be 2-MONOTONE
Output: v^r
see max_dual() for scipy version
"""
x0 = array([1./shape(cap_array)[0] for i in range(shape(cap_array)[0])])
# print x0
Aeq = ones(shape(cap_array)[0])
beq = Budg
A = diag(-ones(shape(cap_array)[0]))
b = zeros(shape(cap_array)[0])
vmax=gm
# vmax = array([chq.Choquet(array(chq.max_choquet(v,Budg)),v) for v in cap_array])
def fprime(x,cap_array,Budg):
cap_int = dot(cap_array.T,x)
x_int = array(max_choquet(cap_int,fx,dfx,Budg))
ch_xr = array([Choquet(x_int,v,fx) for v in cap_array])
ch_diffs = ch_xr - vmax
return ch_diffs
def objfunc(x,cap_array,Budg):
cap_int = dot(cap_array.T,x)
x_int = array(max_choquet(cap_int,fx,dfx,Budg))
ch_xr = array([Choquet(x_int,v,fx) for v in cap_array])
ch_diffs = ch_xr - vmax
# print dot(x, ch_diffs)
return dot(x, ch_diffs)
objfunc_w = lambda x: objfunc(x,cap_array,Budg)
fprime_w = lambda x: fprime(x,cap_array,Budg)
p = NSP(objfunc_w, x0, df=fprime_w, A=A, b=b, Aeq=Aeq, beq=beq, iprint = 25, maxIter = 300, maxFunEvals = 1e7, diffint=1e-10, xtol=1e-9,ftol=1e-7,gtol=1e-5)
r = p.solve('ralg')
print r.xf, r.ff
cap_mix = dot(cap_array.T, array(r.xf))
return r.ff,cap_mix
def __solver__(self, p):
f = lambda x: max(p.f(x))
def df(x):
F = p.f(x)
ind = argmax(F)
return p.df(x, ind)
def iterfcn(*args, **kwargs):
p2.primalIterFcn(*args, **kwargs)
p.xk = p2.xk.copy()
p.fk = p2.fk
p.rk = p2.rk
p.istop = p2.istop
if p.istop and p2.rk <= p2.contol:
p.msg = p2.msg
p.iterfcn()
p2 = NSP(f, p.x0, df=df, xtol = p.xtol, ftol = p.ftol, gtol = p.gtol,\
A=p.A, b=p.b, Aeq=p.Aeq, beq=p.beq, lb=p.lb, ub=p.ub, \
maxFunEvals = p.maxFunEvals, fEnough = p.fEnough, maxIter=p.maxIter, iprint = -1, \
maxtime = p.maxTime, maxCPUTime = p.maxCPUTime, noise = p.noise)
if p.userProvided.c:
p2.c, p2.dc = p.c, p.dc
if p.userProvided.h:
p2.h, p2.dh = p.h, p.dh
p2.primalIterFcn, p2.iterfcn = p2.iterfcn, iterfcn
r2 = p2.solve('ralg')
xf = r2.xf
p.xk = p.xf = xf
p.fk = p.ff = max(p.f(xf))
def __solver__(self, p):
f = lambda x: max(p.f(x))
def df(x):
F = p.f(x)
ind = argmax(F)
return p.df(x, ind)
def iterfcn(*args, **kwargs):
p2.primalIterFcn(*args, **kwargs)
p.xk = p2.xk.copy()
p.fk = p2.fk
p.rk = p2.rk
p.istop = p2.istop
if p.istop and p2.rk <= p2.contol:
p.msg = p2.msg
p.iterfcn()
p2 = NSP(f, p.x0, df=df, xtol = p.xtol, ftol = p.ftol, gtol = p.gtol,\
A=p.A, b=p.b, Aeq=p.Aeq, beq=p.beq, lb=p.lb, ub=p.ub, \
maxFunEvals = p.maxFunEvals, fEnough = p.fEnough, maxIter=p.maxIter, iprint = -1, \
maxtime = p.maxTime, maxCPUTime = p.maxCPUTime, noise = p.noise)
if p.userProvided.c:
p2.c, p2.dc = p.c, p.dc
if p.userProvided.h:
p2.h, p2.dh = p.h, p.dh
p2.primalIterFcn, p2.iterfcn = p2.iterfcn, iterfcn
r2 = p2.solve('ralg')
xf = r2.xf
p.xk = p.xf = xf
p.fk = p.ff = max(p.f(xf))
# x_opt, y_opt, z_opt, t_opt = x(optPoint), y(optPoint), z(optPoint), t(optPoint)
# Example for NSP
# |x[0]| + 1.2*|x[1]| + 1.44*|x[2]| + ... + 1.2^74*|x[74]| + |y-15| +|y+15| + y^2 + -> min
# with some constraints
# coded in FuncDesigner
# all 1st derivatives are obtained via Automatic Differentiation,
# not to be confused with finite-difference derivatives approximation
from numpy import cos, arange
from FuncDesigner import *
from openopt import NSP
x, y = oovars('x y')
N = 75
koeffs = arange(1, N + 1)**1.2 # 1, 1.2, 1.44, ..., 1.2^m, ..., 1.2^N
objective = sum(abs(x) * koeffs) + abs(y - 15) + abs(y + 15) + y**2
constraints = [(y - 1)**2 < 1, abs(y) < 0.5]
constraints.append(
(x - 0.01 * arange(N))**2 < 0.1 * arange(1, N + 1)
) # (x_0-0)**2 < 0.1, (x_1-0.01)**2 < 0.2, (x_2-0.02)**2 < 0.3,...
startPoint = {x: cos(1 + arange(N)), y: 80}
p = NSP(objective, startPoint, maxIter=1e5, constraints=constraints)
r = p.solve('ralg')
x_opt, y_opt = x(r), y(r)
print(max(abs(x_opt)), y_opt)
return False
asa = lambda x: asarray(x).reshape(-1, 1)
solvers = ['ralg', 'amsg2p', 'gsubg']
solvers = [oosolver('amsg2p', gamma=2.0)]
Colors = ['r', 'k', 'b']
lines = []
R = {}
for Tol_p in range(-10, -31, -1):
#print('Tol = 10^%d' % Tol_p)
for i, solver in enumerate(solvers):
p = NSP(obj,
startPoint,
maxIter=1700,
name='Rzhevsky6 (nVars: ' + str(n) + ')',
maxTime=300,
maxFunEvals=1e7,
color=Colors[i])
p.fOpt = 0.0
p.fTol = 10**Tol_p
cb.TMP = 1000
cb.stat = {'dist': [], 'f': [], 'df': []}
r = p.solve(solver, iprint=-1, xtol=0, ftol=0, callback=cb)
R[solver] = hstack(
(asa(cb.stat['dist']), asa(cb.stat['f']), asa(cb.stat['df'])))
print('itn df dx %d %0.1e %0.1e' %
(-r.evals['df'], r.ff - p.fOpt, norm(p.xk - 1.0)))
# print('objective evals: %d gradient evals: %d ' % (r.evals['f'],r.evals['df']))
# print('distance to f*: %0.1e' % (r.ff-p.fOpt))
# print('distance to x*: %0.1e' % norm(p.xk-1.0))
xOpt = [-1.26257e-1, -3.43781e-2, -6.85720e-3, 2.63607e-2, 6.72950e-2, -2.78399e-1, 7.42187e-2, 1.38524e-1, 8.40313e-2, 3.85803e-2]
def cb(p):
tmp = ceil(log10(norm(xOpt - p.xk)))
if tmp < cb.TMP:
# print 'distance:', tmp, 'itn:', p.iter, 'n_func:', p.nEvals['f'], 'n_grad:', -p.nEvals['df']
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x:asarray(x).reshape(-1, 1)
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj, startPoint, maxIter = 17000, name = 'Rzhevsky1 (nVars: ' + str(n)+')', maxTime = 300, maxFunEvals=1e7, color = Colors[i])
p.fTol = 0.5e-10
p.fEnough = -0.84140833459
p.fOpt = -0.841408334596
cb.TMP = 1000
cb.stat = {'dist':[], 'f':[], 'df':[]}
r = p.solve(solver, iprint=10, ftol = 1e-15, xtol = 1e-10, debug=0, show = solver == solvers[-1], plot = 0, callback=cb)
R[solver] = hstack((asa(cb.stat['dist']), asa(cb.stat['f']), asa(cb.stat['df'])))
'''
--------------------------------------------------
solver: gsubg problem: rjevsky1 (nVars: 10) type: NSP goal: minimum
iter objFunVal
0 1.573e+02
10 -7.780e-01
20 -8.410e-01
if __name__ == '__main__':
import sys, os.path as pth
sys.path.insert(
0,
pth.split(
pth.split(
pth.split(pth.split(pth.realpath(
pth.dirname(__file__)))[0])[0])[0])[0])
from openopt import NSP
from numpy import ones, arange
N = 4
p = NSP(lambda x: abs(x).sum(), ones([N, 1]), maxFunEvals=1e6, plot=False)
p.r0 = N
r = p.solve('ShorEllipsoid')
from numpy import diag, ones, inf, any, copy, sqrt
from openopt.kernel.baseSolver import baseSolver
class ShorEllipsoid(baseSolver):
__name__ = 'ShorEllipsoid'
__license__ = "BSD"
__authors__ = "Dmitrey"
__alg__ = "Naum Z. Shor modificated method of ellipsoids"
iterfcnConnected = True
#__optionalDataThatCanBeHandled__ = ['A', 'Aeq', 'b', 'beq', 'lb', 'ub', 'c', 'h']
def __init__(self):
pass
cb.stat["f"].append(p.nEvals["f"])
cb.stat["df"].append(-p.nEvals["df"])
return False
asa = lambda x: asarray(x).reshape(-1, 1)
lines = []
R = {}
for i, solver in enumerate(solvers):
p = NSP(
obj,
startPoint,
maxIter=1700,
name="Rzhevsky5 (nVars: " + str(n) + ")",
maxTime=300,
maxFunEvals=1e7,
color=Colors[i],
)
p.fTol = 0.5e-10
cb.TMP = 1000
cb.stat = {"dist": [], "f": [], "df": []}
p.fOpt = -34.408608965509742
p.maxTime = 30
r = p.solve(solver, iprint=1, xtol=1e-15, ftol=1e-15, gtol=1e-15, show=solver == solvers[-1], plot=0, callback=cb)
R[solver] = hstack((asa(cb.stat["dist"]), asa(cb.stat["f"]), asa(cb.stat["df"])))
"""
solver: gsubg problem: rjevsky5 (nVars: 50) type: NSP goal: minimum
iter objFunVal
# or
# return 0
from openopt import NSP
from numpy import cos, asfarray, arange, sign
N = 75
f = lambda x: sum(1.2 ** arange(len(x)) * abs(x))
df = lambda x: 1.2 ** arange(len(x)) * sign(x)
x0 = cos(1 + asfarray(range(N)))
# non-linear constraint c(x) <= 0:
c = lambda x: abs(x[4] - 0.8) + abs(x[5] - 1.5) - 0.015
p = NSP(f, x0, df=df, c=c, callback=MyIterFcn, contol=1e-5, maxIter=1e4, iprint=100, xtol=1e-8, ftol=1e-8)
# optional:
# p.plot = 1
r = p.solve("ralg")
print r.xf[:8]
"""
-----------------------------------------------------
solver: ralg problem: unnamed goal: minimum
iter objFunVal log10(maxResidual)
0 2.825e+06 0.02
--= non-feasible ralg iter =--
itn: 0
curr f: [ 2824966.83813157]
max constraint value 1.04116752789
x, y = oovars('x y')
N = 75
koeffs = arange(1, N + 1)**1.2 # 1, 1.2, 1.44, ..., 1.2^m, ..., 1.2^N
objective = sum(abs(x) * koeffs) + abs(y - 15) + abs(y + 15) + y**2
constraints = [(y - 1)**2 < 1,
abs(y) < 0.5,
abs(x[0]) < 1e-5,
abs(x[N - 1]) < 1e-5]
constraints.append(
(x - 0.01 * arange(N))**2 < 0.1 * arange(1, N + 1)
) # (x_0-0)**2 < 0.1, (x_1-0.01)**2 < 0.2, (x_2-0.02)**2 < 0.2,...
startPoint = {x: cos(1 + arange(N)), y: 80}
p = NSP(objective, startPoint, maxIter=1e5, constraints=constraints)
r = p.solve('ralg')
x_opt, y_opt = x(r), y(r)
print(max(abs(x_opt)), y_opt)
'''
expected output:
[...]
876 3.004e+01 -100.00
istop: 4 (|| F[k] - F[k-1] || < ftol)
Solver: Time Elapsed = 7.98 CPU Time Elapsed = 7.97
objFunValue: 30.042539 (feasible, MaxResidual = 0)
(6.6277698279489041e-06, 0.20306221768582972)
'''
print 'start point: f1 = %e f2 = %e' % (f1(startPoint), f2(startPoint))
#print "start point: norm(f1') = %e norm(f2') = %e" % (norm(f1.D(startPoint, y)), norm(f2.D(startPoint, x)))
ralg = oosolver('ralg')
gsubg = oosolver('gsubg', addASG = False)
solvers = [ralg]
#solvers = [oosolver('gsubg', zhurb = 20, dual=False)]
solvers = ['ipopt', gsubg, 'scipy_cg']
solvers = [gsubg]
#solvers = ['ipopt']
#solvers = ['slmvm2']
#solvers = ['slmvm1']
#solvers = ['mma']
Colors = ['r', 'k','b']
lines = []
for i, solver in enumerate(solvers):
p = NSP(f, startPoint, maxIter = 300, name = 'ns' + str(N+K), maxTime = 15000, maxFunEvals=1e7, color = Colors[i])
#p.constraints = y>-100
p.fEnough = 1.0e-1#e-1
p.fTol = 0.5e-1
p.debug = 1
#p.constraints = (y > 5)(tol=1e-4) #x>1e-1 #[2*y<sin(arange(N))]
#r = p.manage(solver, iprint=10, xtol = 1e-9, ftol = 1e-9, show = solver == solvers[-1], maxIter = 10000)
r = p.solve(solver, iprint=10, xtol = 1e-6, ftol = 1e-6, debug=0, show = solver == solvers[-1], plot = 0)
print 'end point: f1 = %e f2 = %e' % (f1(r.xf), f2(r.xf))
from openopt import NSP
from numpy import cos, asfarray, arange, sign
N = 75
f = lambda x: sum(1.2**arange(len(x)) * abs(x))
df = lambda x: 1.2**arange(len(x)) * sign(x)
x0 = cos(1 + asfarray(range(N)))
#non-linear constraint c(x) <= 0:
c = lambda x: abs(x[4] - 0.8) + abs(x[5] - 1.5) - 0.015
p = NSP(f,
x0,
df=df,
c=c,
callback=MyIterFcn,
contol=1e-5,
maxIter=1e4,
iprint=100,
xtol=1e-8,
ftol=1e-8)
#optional:
#p.plot = 1
r = p.solve('ralg')
print r.xf[:8]
"""
-----------------------------------------------------
solver: ralg problem: unnamed goal: minimum
iter objFunVal log10(maxResidual)
0 2.825e+06 0.02
--= non-feasible ralg iter =--
|x1| + 1.2|x2| + 1.44|x3| + ... + 1.2^N |xN| -> min N=75 x0 = [cos(1), cos(2), ..., cos(N)] x_opt = all-zeros f_opt = 0 """ from numpy import * from openopt import NSP N = 75 objFun = lambda x: sum(1.2 ** arange(len(x)) * abs(x)) x0 = cos(1+asfarray(range(N))) p = NSP(objFun, x0, maxFunEvals = 1e7, xtol = 1e-8) #These assignments are also correct: #p = NLP(objFun, x0=x0) #p = NLP(f=objFun, x0=x0) #p = NLP(ftol = 1e-5, contol = 1e-5, f=objFun, x0=x0) p.maxIter = 5000 #optional (requires matplotlib installed) #p.plot = 1 #p.graphics.xlabel = 'cputime'#default: time, case-unsensetive; also here maybe 'cputime', 'niter' #OPTIONAL: user-supplied gradient/subgradient p.df = lambda x: 1.2 ** arange(len(x)) * sign(x)
# print 'distance:', tmp, 'itn:', p.iter, 'n_func:', p.nEvals['f'], 'n_grad:', -p.nEvals['df']
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x:asarray(x).reshape(-1, 1)
solvers = ['ralg', 'amsg2p', 'gsubg']
solvers = [oosolver('amsg2p', gamma = 2.0)]
Colors = ['r', 'k','b']
lines = []
R = {}
for Tol_p in range(-10, -31, -1):
#print('Tol = 10^%d' % Tol_p)
for i, solver in enumerate(solvers):
p = NSP(obj, startPoint, maxIter = 1700, name = 'Rzhevsky6 (nVars: ' + str(n)+')', maxTime = 300, maxFunEvals=1e7, color = Colors[i])
p.fOpt = 0.0
p.fTol = 10**Tol_p
cb.TMP = 1000
cb.stat = {'dist':[], 'f':[], 'df':[]}
r = p.solve(solver, iprint=-1, xtol = 0, ftol = 0, callback = cb)
R[solver] = hstack((asa(cb.stat['dist']), asa(cb.stat['f']), asa(cb.stat['df'])))
print('itn df dx %d %0.1e %0.1e' % (-r.evals['df'], r.ff-p.fOpt, norm(p.xk-1.0)))
# print('objective evals: %d gradient evals: %d ' % (r.evals['f'],r.evals['df']))
# print('distance to f*: %0.1e' % (r.ff-p.fOpt))
# print('distance to x*: %0.1e' % norm(p.xk-1.0))
'''
--------------------------------------------------
solver: gsubg problem: rjevsky6 (nVars: 15) type: NSP goal: minimum
iter objFunVal
0 2.408e+02
from numpy import arange
from numpy.linalg import norm
from openopt import NSP, oosolver
from FuncDesigner import *
N = 100000
x = oovar('x')
startPoint = {x: 1 + 1.0 / arange(1, N+1)}
S = 1e4 ** (1.0/arange(1, N+1))
arr = sin(arange(N))
f = sum((x-arr)**2 * S) / 1e4
solvers = [oosolver('ralg')]
solvers = [oosolver('gsubg', dual=True, zhurb = 50)]
#solvers = [oosolver('gsubg', zhurb = 20, dual=False)]
#solvers = ['ipopt']
#solvers = ['slmvm2']
#solvers = ['mma']
for solver in solvers:
p = NSP(f, startPoint, maxIter = 10000, maxTime = 15000, maxFunEvals=1e7)
p.fEnough = 1.5e-1
p.fTol = 5.0e-1
#p.constraints = (y > 5)(tol=1e-4) #x>1e-1 #[2*y<sin(arange(N))]
#r = p.solve(solver, iprint=10, xtol = 1e-36, ftol = 1e-16, show = solver == solvers[-1])
r = p.manage(solver, iprint=1, xtol = 1e-8, ftol = 1e-7, show = solver == solvers[-1])
if __name__ == '__main__':
import sys, os.path as pth
sys.path.insert(0,pth.split(pth.split(pth.split(pth.split(pth.realpath(pth.dirname(__file__)))[0])[0])[0])[0])
from openopt import NSP
from numpy import ones, arange
N = 4
p = NSP(lambda x: abs(x).sum(), ones([N,1]), maxFunEvals = 1e6, plot = False)
p.r0 = N
r = p.solve('ShorEllipsoid')
from numpy import diag, ones, inf, any, copy, sqrt
from openopt.kernel.baseSolver import baseSolver
class ShorEllipsoid(baseSolver):
__name__ = 'ShorEllipsoid'
__license__ = "BSD"
__authors__ = "Dmitrey"
__alg__ = "Naum Z. Shor modificated method of ellipsoids"
iterfcnConnected = True
#__optionalDataThatCanBeHandled__ = ['A', 'Aeq', 'b', 'beq', 'lb', 'ub', 'c', 'h']
def __init__(self): pass
def __solver__(self, p):
#if p.useScaling: p.err('ShorEllipsoid + scaling isn''t implemented yet')
n = p.n
B = diag(ones(n))
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x: asarray(x).reshape(-1, 1)
lines = []
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj,
startPoint,
maxIter=1700,
name='Rzhevsky5 (nVars: ' + str(n) + ')',
maxTime=300,
maxFunEvals=1e7,
color=Colors[i])
p.fTol = 0.5e-10
cb.TMP = 1000
cb.stat = {'dist': [], 'f': [], 'df': []}
p.fOpt = -34.408608965509742
p.maxTime = 30
r = p.solve(solver,
iprint=1,
xtol=1e-15,
ftol=1e-15,
gtol=1e-15,
show=solver == solvers[-1],
plot=0,
def __solver__(self, p):
n = p.n
x0 = copy(p.x0)
xPrev = x0.copy()
xf = x0.copy()
xk = x0.copy()
p.xk = x0.copy()
f0 = p.f(x0)
fk = f0
ff = f0
p.fk = fk
df0 = p.df(x0)
#####################################################################
## #handling box-bounded problems
## if p.__isNoMoreThanBoxBounded__():
## for k in range(int(p.maxIter)):
##
## #end of handling box-bounded problems
isBB = p.__isNoMoreThanBoxBounded__()
## isBB = 0
H = diag(ones(p.n))
if not p.userProvided.c:
p.c = lambda x: array([])
p.dc = lambda x: array([]).reshape(0, p.n)
if not p.userProvided.h:
p.h = lambda x: array([])
p.dh = lambda x: array([]).reshape(0, p.n)
p.use_subproblem = 'QP'
#p.use_subproblem = 'LLSP'
for k in range(p.maxIter + 4):
if isBB:
f0 = p.f(xk)
df = p.df(xk)
direction = -df
f1 = p.f(xk + direction)
ind_l = direction <= p.lb - xk
direction[ind_l] = (p.lb - xk)[ind_l]
ind_u = direction >= p.ub - xk
direction[ind_u] = (p.ub - xk)[ind_u]
ff = p.f(xk + direction)
## print 'f0', f0, 'f1', f1, 'ff', ff
else:
mr = p.getMaxResidual(xk)
if mr > p.contol: mr_grad = p.getMaxConstrGradient(xk)
lb = p.lb - xk #- p.contol/2
ub = p.ub - xk #+ p.contol/2
c, dc, h, dh, df = p.c(xk), p.dc(xk), p.h(xk), p.dh(xk), p.df(
xk)
A, Aeq = vstack((dc, p.A)), vstack((dh, p.Aeq))
b = concatenate((-c, p.b - p.matmult(p.A, xk))) #+ p.contol/2
beq = concatenate((-h, p.beq - p.matmult(p.Aeq, xk)))
if b.size != 0:
isFinite = isfinite(b)
ind = where(isFinite)[0]
A, b = A[ind], b[ind]
if beq.size != 0:
isFinite = isfinite(beq)
ind = where(isFinite)[0]
Aeq, beq = Aeq[ind], beq[ind]
if p.use_subproblem == 'LP': #linear
linprob = LP(df, A=A, Aeq=Aeq, b=b, beq=beq, lb=lb, ub=ub)
linprob.iprint = -1
r2 = linprob.solve(
'cvxopt_glpk') # TODO: replace lpSolve by autoselect
if r2.istop <= 0:
p.istop = -12
p.msg = "failed to solve LP subproblem"
return
elif p.use_subproblem == 'QP': #quadratic
qp = QP(H=H,
f=df,
A=A,
Aeq=Aeq,
b=b,
beq=beq,
lb=lb,
ub=ub)
qp.iprint = -1
r2 = qp.solve(
'cvxopt_qp') # TODO: replace solver by autoselect
#r2 = qp.solve('qld') # TODO: replace solver by autoselect
if r2.istop <= 0:
for i in range(4):
if p.debug:
p.warn("iter " + str(k) + ": attempt Num " +
str(i) +
" to solve QP subproblem has failed")
#qp.f += 2*N*sum(qp.A,0)
A2 = vstack((A, Aeq, -Aeq))
b2 = concatenate(
(b, beq, -beq)) + pow(10, i) * p.contol
qp = QP(H=H, f=df, A=A2, b=b2, iprint=-5)
qp.lb = lb - pow(10, i) * p.contol
qp.ub = ub + pow(10, i) * p.contol
# I guess lb and ub don't matter here
try:
r2 = qp.solve(
'cvxopt_qp'
) # TODO: replace solver by autoselect
except:
r2.istop = -11
if r2.istop > 0: break
if r2.istop <= 0:
p.istop = -11
p.msg = "failed to solve QP subproblem"
return
elif p.use_subproblem == 'LLSP':
direction_c = getConstrDirection(p,
xk,
regularization=1e-7)
else:
p.err('incorrect or unknown subproblem')
if isBB:
X0 = xk.copy()
N = 0
result, newX = chLineSearch(p, X0, direction, N, isBB)
elif p.use_subproblem != 'LLSP':
duals = r2.duals
N = 1.05 * abs(duals).sum()
direction = r2.xf
X0 = xk.copy()
result, newX = chLineSearch(p, X0, direction, N, isBB)
else: # case LLSP
direction_f = -df
p2 = NSP(LLSsubprobF, [0.8, 0.8],
ftol=0,
gtol=0,
xtol=1e-5,
iprint=-1)
p2.args.f = (xk, direction_f, direction_c, p, 1e20)
r_subprob = p2.solve('ralg')
alpha = r_subprob.xf
newX = xk + alpha[0] * direction_f + alpha[1] * direction_c
# dw = (direction_f * direction_c).sum()
# cos_phi = dw/p.norm(direction_f)/p.norm(direction_c)
# res_0, res_1 = p.getMaxResidual(xk), p.getMaxResidual(xk+1e-1*direction_c)
# print cos_phi, res_0-res_1
# res_0 = p.getMaxResidual(xk)
# optimConstrPoint = getDirectionOptimPoint(p, p.getMaxResidual, xk, direction_c)
# res_1 = p.getMaxResidual(optimConstrPoint)
#
# maxConstrLimit = p.contol
#xk = getDirectionOptimPoint(p, p.f, optimConstrPoint, -optimConstrPoint+xk+direction_f, maxConstrLimit = maxConstrLimit)
#print 'res_0', res_0, 'res_1', res_1, 'res_2', p.getMaxResidual(xk)
#xk = getDirectionOptimPoint(p, p.f, xk, direction_f, maxConstrLimit)
#newX = xk.copy()
result = 0
# x_0 = X0.copy()
# N = j = 0
# while p.getMaxResidual(x_0) > Residual0 + 0.1*p.contol:
# j += 1
# x_0 = xk + 0.75**j * (X0-xk)
# X0 = x_0
# result, newX = 0, X0
# print 'newIterResidual = ', p.getMaxResidual(x_0)
if result != 0:
p.istop = result
p.xf = newX
return
xk = newX.copy()
fk = p.f(xk)
p.xk, p.fk = copy(xk), copy(fk)
#p._df = p.df(xk)
####################
p.iterfcn()
if p.istop:
p.xf = xk
p.ff = fk
#p._df = g FIXME: implement me
return
def __solver__(self, p):
if SMALL_DELTA_X in p.kernelIterFuncs.keys(): p.kernelIterFuncs.pop(SMALL_DELTA_X)
if SMALL_DELTA_F in p.kernelIterFuncs.keys(): p.kernelIterFuncs.pop(SMALL_DELTA_F)
if self.nspSolver == 'autoselect':
nspSolver = 'amsg2p' if p.isUC else 'ralg'
else:
nspSolver = self.nspSolver
# nspSolver = 'ralg'
way = 3 if nspSolver == 'ralg' else 2
if way == 1:
use2 = False
f = lambda x: sum(abs(p.f(x)))
def df(x):
return dot(p.df(x), sign(p.f(x)))
elif way == 2:
use2 = True
f = lambda x: sum(p.f(x)**2)
def df(x):
return 2.0*dot(p.f(x), p.df(x))
elif way == 3:
use2 = False
f = lambda x: max(abs(p.f(x)))
def df(x):
F = p.f(x)
ind = argmax(abs(F))
return p.df(x, ind) * sign(F[ind])
FTOL = p.ftol**2 if use2 else p.ftol
def iterfcn(*args, **kwargs):
p2.primalIterFcn(*args, **kwargs)
p.xk = p2.xk.copy()
Fk = norm(p.f(p.xk), inf)
p.rk = p.getMaxResidual(p.xk)
#TODO: ADD p.rk
if p.nEvals['f'] > p.maxFunEvals:
p.istop = p2.istop = IS_MAX_FUN_EVALS_REACHED
elif p2.istop!=0:
if Fk < FTOL and p.rk < p.contol:
p.istop = 15
msg_contol = '' if p.isUC else 'and contol '
p.msg = 'solution with required ftol ' + msg_contol+ 'has been reached'
else: p.istop = p2.istop
p.iterfcn()
return p.istop
p2 = NSP(f, p.x0, df=df, xtol = p.xtol/1e16, gtol = p.gtol/1e16,\
A=p.A, b=p.b, Aeq=p.Aeq, beq=p.beq, lb=p.lb, ub=p.ub, \
maxFunEvals = p.maxFunEvals, fEnough = FTOL, maxIter=p.maxIter, iprint = -1, \
maxtime = p.maxTime, maxCPUTime = p.maxCPUTime, noise = p.noise, fOpt = 0.0)
if p.userProvided.c:
p2.c, p2.dc = p.c, p.dc
if p.userProvided.h:
p2.h, p2.dh = p.h, p.dh
p2.primalIterFcn, p2.iterfcn = p2.iterfcn, iterfcn
if p.debug: p2.iprint = 1
if nspSolver == 'ralg':
if p.isUC: p2.ftol = p.ftol / 1e16
else: p2.ftol = 0.0
else:
p2.ftol = 0.0
p2.xtol = 0.0
p2.gtol = 0.0
if use2:
p2.fTol = 0.5*p.ftol ** 2
else:
p2.fTol = 0.5*p.ftol
r2 = p2.solve(nspSolver)
#xf = fsolve(p.f, p.x0, fprime=p.df, xtol = p.xtol, maxfev = p.maxFunEvals)
xf = r2.xf
p.xk = p.xf = xf
p.fk = p.ff = asfarray(norm(p.f(xf), inf)).flatten()
solvers = ['ralg', 'amsg2p']
solvers = ['gsubg']
Colors = ['r', 'k','b']
xOpt = 1.0/n
def cb(p):
tmp = ceil(log10(norm(xOpt - p.xk)))
if tmp < cb.TMP:
# print 'distance:', tmp, 'itn:', p.iter, 'n_func:', p.nEvals['f'], 'n_grad:', -p.nEvals['df']
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x:asarray(x).reshape(-1, 1)
R = {}
lines = []
for i, solver in enumerate(solvers):
p = NSP(obj, startPoint, maxIter = 4700, name = 'Rzhevsky3 (nVars: ' + str(n)+')', maxTime = 30000, maxFunEvals=1e7, color = Colors[i])
#p.maxIter = 10#; p.useSparse = False
p.fEnough = 1.0e-5
p.fOpt = 1.0e-5
p.fTol = 0.5e-5
cb.TMP = 1000
cb.stat = {'dist':[], 'f':[], 'df':[]}
r = p.manage(solver, iprint=1, xtol = 1e-10, ftol = 1e-10, show = solver == solvers[-1], plot = 0, callback = cb)
R[solver] = hstack((asa(cb.stat['dist']), asa(cb.stat['f']), asa(cb.stat['df'])))
if tmp < cb.TMP:
# print 'distance:', tmp, 'itn:', p.iter, 'n_func:', p.nEvals['f'], 'n_grad:', -p.nEvals['df']
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x: asarray(x).reshape(-1, 1)
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj,
startPoint,
maxIter=17000,
name='Rzhevsky1 (nVars: ' + str(n) + ')',
maxTime=300,
maxFunEvals=1e7,
color=Colors[i])
p.fTol = 0.5e-10
p.fEnough = -0.84140833459
p.fOpt = -0.841408334596
cb.TMP = 1000
cb.stat = {'dist': [], 'f': [], 'df': []}
r = p.solve(solver,
iprint=10,
ftol=1e-15,
xtol=1e-10,
debug=0,
show=solver == solvers[-1],
plot=0,
from numpy import cos, arange
from FuncDesigner import *
from openopt import NSP
x, y = oovars('x y')
N = 75
koeffs = arange(1, N+1) ** 1.2 # 1, 1.2, 1.44, ..., 1.2^m, ..., 1.2^N
objective = sum(abs(x) * koeffs) + abs(y-15) + abs(y+15) + y**2
constraints = [(y-1)**2<1, abs(y) < 0.5, abs(x[0]) < 1e-5, abs(x[N-1]) < 1e-5]
constraints.append((x - 0.01*arange(N))**2 < 0.1*arange(1, N+1)) # (x_0-0)**2 < 0.1, (x_1-0.01)**2 < 0.2, (x_2-0.02)**2 < 0.2,...
startPoint = {x: cos(1+arange(N)), y:80}
p = NSP(objective, startPoint, maxIter = 1e5, constraints = constraints)
r = p.solve('ralg')
x_opt, y_opt = x(r), y(r)
print(max(abs(x_opt)), y_opt)
'''
expected output:
[...]
876 3.004e+01 -100.00
istop: 4 (|| F[k] - F[k-1] || < ftol)
Solver: Time Elapsed = 7.98 CPU Time Elapsed = 7.97
objFunValue: 30.042539 (feasible, MaxResidual = 0)
(6.6277698279489041e-06, 0.20306221768582972)
'''
def cb(p):
tmp = ceil(log10(norm(asarray(X) - hstack((X[0], p.xk, X[-1])))))
if tmp < cb.TMP:
# print 'distance:', tmp, 'itn:', p.iter, 'n_func:', p.nEvals['f'], 'n_grad:', -p.nEvals['df']
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x:asarray(x).reshape(-1, 1)
Colors = ['r', 'k','b']
lines = []
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj, startPoint, fixedVars=(x[0], x[-1]), maxTime = 20, name = 'Rzhevsky7 (nVars: ' + str(n)+')', maxFunEvals=1e7, color = Colors[i])
p._prepare()
p.c=None
#p.fEnough = 2.08983385058799+4e-10
p.fOpt = obj(T_optPoint)
p.fTol = 0.5e-15
cb.TMP = 1000
cb.stat = {'dist':[], 'f':[], 'df':[]}
r = p.solve(solver, iprint=10, xtol = 1e-10, ftol = 1e-16, gtol = 1e-10, debug=0, show = solver == solvers[-1], plot = 0, callback = cb)
print('objective evals: %d gradient evals: %d ' % (r.evals['f'],r.evals['df']))
print('distance to f*: %0.1e' % (r.ff-p.fOpt))
print('distance to x*: %0.1e' % (norm(asarray(X) - hstack((X[0], p.xk, X[-1])))))
R[solver] = hstack((asa(cb.stat['dist']), asa(cb.stat['f']), asa(cb.stat['df'])))
'''
--------------------------------------------------
solver: gsubg problem: rjevsky6 (nVars: 30) type: NSP goal: minimum
xOpt = [-1.26257e-1, -3.43781e-2, -6.85720e-3, 2.63607e-2, 6.72950e-2, -2.78399e-1, 7.42187e-2, 1.38524e-1, 8.40313e-2, 3.85803e-2, -0.841408334596]
def cb(p):
tmp = ceil(log10(norm(xOpt - p.xk)))
if tmp < cb.TMP:
# print 'distance:', tmp, 'itn:', p.iter, 'n_func:', p.nEvals['f'], 'n_grad:', -p.nEvals['df']
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x:asarray(x).reshape(-1, 1)
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj, startPoint, maxIter = 17000, name = 'Rzhevsky2 (nVars: ' + str(n)+')', maxTime = 300, maxFunEvals=1e7, color = Colors[i])
p.fTol = 0.5e-12
p.maxTime = 100
#p.fEnough = -0.841408334595
p.fOpt = -0.841408334596
cb.TMP = 1000
cb.stat = {'dist':[], 'f':[], 'df':[]}
r = p.solve(solver, iprint=10, ftol = 1e-15, xtol = 1e-12, show = solver == solvers[-1], plot = 0, callback = cb)
R[solver] = hstack((asa(cb.stat['dist']), asa(cb.stat['f']), asa(cb.stat['df'])))
'''
--------------------------------------------------
solver: gsubg problem: rjevsky2 (nVars: 10) type: NSP goal: minimum
iter objFunVal
0 5.337e+03
10 -5.386e-01
20 -8.399e-01
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x: asarray(x).reshape(-1, 1)
Colors = ['r', 'k', 'b']
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj,
startPoint,
maxIter=1700,
name='Rzhevsky4 (nVars: ' + str(n) + ')',
maxTime=300,
maxFunEvals=1e7,
color=Colors[i])
p.fTol = 0.5e-20
p.fOpt = 0.0
cb.TMP = 1000
cb.stat = {'dist': [], 'f': [], 'df': []}
r = p.solve(solver,
iprint=10,
xtol=1e-20,
ftol=1e-20,
debug=0,
show=solver == solvers[-1],
plot=0,
callback=cb)
tmp = ceil(log10(norm(xOpt - p.xk)))
if tmp < cb.TMP:
# print 'distance:', tmp, 'itn:', p.iter, 'n_func:', p.nEvals['f'], 'n_grad:', -p.nEvals['df']
cb.TMP = tmp
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x:asarray(x).reshape(-1, 1)
Colors = ['r', 'k','b']
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj, startPoint, maxIter = 1700, name = 'Rzhevsky4 (nVars: ' + str(n)+')', maxTime = 300, maxFunEvals=1e7, color = Colors[i])
p.fTol = 0.5e-20
p.fOpt = 0.0
cb.TMP = 1000
cb.stat = {'dist':[], 'f':[], 'df':[]}
r = p.solve(solver, iprint=10, xtol = 1e-20, ftol = 1e-20, debug=0, show = solver == solvers[-1], plot = 0, callback = cb)
R[solver] = hstack((asa(cb.stat['dist']), asa(cb.stat['f']), asa(cb.stat['df'])))
'''
--------------------------------------------------
solver: gsubg problem: rjevsky3 (nVars: 10) type: NSP goal: minimum
iter objFunVal
0 9.789e-01
10 5.155e-03
12 2.168e-03
istop: 16 (optimal solution wrt required fTol has been obtained)
Solver: Time Elapsed = 1.1 CPU Time Elapsed = 1.1
def __solver__(self, p):
n = p.n
x0 = copy(p.x0)
xPrev = x0.copy()
xf = x0.copy()
xk = x0.copy()
p.xk = x0.copy()
f0 = p.f(x0)
fk = f0
ff = f0
p.fk = fk
df0 = p.df(x0)
#####################################################################
## #handling box-bounded problems
## if p.__isNoMoreThanBoxBounded__():
## for k in range(int(p.maxIter)):
##
## #end of handling box-bounded problems
isBB = p.__isNoMoreThanBoxBounded__()
## isBB = 0
H = diag(ones(p.n))
if not p.userProvided.c:
p.c = lambda x : array([])
p.dc = lambda x : array([]).reshape(0, p.n)
if not p.userProvided.h:
p.h = lambda x : array([])
p.dh = lambda x : array([]).reshape(0, p.n)
p.use_subproblem = 'QP'
#p.use_subproblem = 'LLSP'
for k in range(p.maxIter+4):
if isBB:
f0 = p.f(xk)
df = p.df(xk)
direction = -df
f1 = p.f(xk+direction)
ind_l = direction<=p.lb-xk
direction[ind_l] = (p.lb-xk)[ind_l]
ind_u = direction>=p.ub-xk
direction[ind_u] = (p.ub-xk)[ind_u]
ff = p.f(xk + direction)
## print 'f0', f0, 'f1', f1, 'ff', ff
else:
mr = p.getMaxResidual(xk)
if mr > p.contol: mr_grad = p.getMaxConstrGradient(xk)
lb = p.lb - xk #- p.contol/2
ub = p.ub - xk #+ p.contol/2
c, dc, h, dh, df = p.c(xk), p.dc(xk), p.h(xk), p.dh(xk), p.df(xk)
A, Aeq = vstack((dc, p.A)), vstack((dh, p.Aeq))
b = concatenate((-c, p.b-p.matmult(p.A,xk))) #+ p.contol/2
beq = concatenate((-h, p.beq-p.matmult(p.Aeq,xk)))
if b.size != 0:
isFinite = isfinite(b)
ind = where(isFinite)[0]
A, b = A[ind], b[ind]
if beq.size != 0:
isFinite = isfinite(beq)
ind = where(isFinite)[0]
Aeq, beq = Aeq[ind], beq[ind]
if p.use_subproblem == 'LP': #linear
linprob = LP(df, A=A, Aeq=Aeq, b=b, beq=beq, lb=lb, ub=ub)
linprob.iprint = -1
r2 = linprob.solve('cvxopt_glpk') # TODO: replace lpSolve by autoselect
if r2.istop <= 0:
p.istop = -12
p.msg = "failed to solve LP subproblem"
return
elif p.use_subproblem == 'QP': #quadratic
qp = QP(H=H,f=df, A=A, Aeq=Aeq, b=b, beq=beq, lb=lb, ub = ub)
qp.iprint = -1
r2 = qp.solve('cvxopt_qp') # TODO: replace solver by autoselect
#r2 = qp.solve('qld') # TODO: replace solver by autoselect
if r2.istop <= 0:
for i in range(4):
if p.debug: p.warn("iter " + str(k) + ": attempt Num " + str(i) + " to solve QP subproblem has failed")
#qp.f += 2*N*sum(qp.A,0)
A2 = vstack((A, Aeq, -Aeq))
b2 = concatenate((b, beq, -beq)) + pow(10,i)*p.contol
qp = QP(H=H,f=df, A=A2, b=b2, iprint = -5)
qp.lb = lb - pow(10,i)*p.contol
qp.ub = ub + pow(10,i)*p.contol
# I guess lb and ub don't matter here
try:
r2 = qp.solve('cvxopt_qp') # TODO: replace solver by autoselect
except:
r2.istop = -11
if r2.istop > 0: break
if r2.istop <= 0:
p.istop = -11
p.msg = "failed to solve QP subproblem"
return
elif p.use_subproblem == 'LLSP':
direction_c = getConstrDirection(p, xk, regularization = 1e-7)
else: p.err('incorrect or unknown subproblem')
if isBB:
X0 = xk.copy()
N = 0
result, newX = chLineSearch(p, X0, direction, N, isBB)
elif p.use_subproblem != 'LLSP':
duals = r2.duals
N = 1.05*abs(duals).sum()
direction = r2.xf
X0 = xk.copy()
result, newX = chLineSearch(p, X0, direction, N, isBB)
else: # case LLSP
direction_f = -df
p2 = NSP(LLSsubprobF, [0.8, 0.8], ftol=0, gtol=0, xtol = 1e-5, iprint = -1)
p2.args.f = (xk, direction_f, direction_c, p, 1e20)
r_subprob = p2.solve('ralg')
alpha = r_subprob.xf
newX = xk + alpha[0]*direction_f + alpha[1]*direction_c
# dw = (direction_f * direction_c).sum()
# cos_phi = dw/p.norm(direction_f)/p.norm(direction_c)
# res_0, res_1 = p.getMaxResidual(xk), p.getMaxResidual(xk+1e-1*direction_c)
# print cos_phi, res_0-res_1
# res_0 = p.getMaxResidual(xk)
# optimConstrPoint = getDirectionOptimPoint(p, p.getMaxResidual, xk, direction_c)
# res_1 = p.getMaxResidual(optimConstrPoint)
#
# maxConstrLimit = p.contol
#xk = getDirectionOptimPoint(p, p.f, optimConstrPoint, -optimConstrPoint+xk+direction_f, maxConstrLimit = maxConstrLimit)
#print 'res_0', res_0, 'res_1', res_1, 'res_2', p.getMaxResidual(xk)
#xk = getDirectionOptimPoint(p, p.f, xk, direction_f, maxConstrLimit)
#newX = xk.copy()
result = 0
# x_0 = X0.copy()
# N = j = 0
# while p.getMaxResidual(x_0) > Residual0 + 0.1*p.contol:
# j += 1
# x_0 = xk + 0.75**j * (X0-xk)
# X0 = x_0
# result, newX = 0, X0
# print 'newIterResidual = ', p.getMaxResidual(x_0)
if result != 0:
p.istop = result
p.xf = newX
return
xk = newX.copy()
fk = p.f(xk)
p.xk, p.fk = copy(xk), copy(fk)
#p._df = p.df(xk)
####################
p.iterfcn()
if p.istop:
p.xf = xk
p.ff = fk
#p._df = g FIXME: implement me
return
from numpy import arange
from numpy.linalg import norm
from openopt import NSP, oosolver
from FuncDesigner import *
N = 300
x = oovar('x')
startPoint = {x: 1 + 1.0 / arange(1, N)}
S = 1e4 ** (1.0/arange(1, N))
#f = abs(x[0]) + S * abs(x[1]) + S**2 * abs(x[2])
f = sum(abs(x)*S)
solvers = [oosolver('ralg')]
solvers = [oosolver('gsubg', addASG = True)]
#solvers = [oosolver('gsubg', zhurb = 20, dual=False)]
#solvers = ['ipopt']
#solvers = ['slmvm2']
#solvers = ['mma']
for solver in solvers:
p = NSP(f, startPoint, maxIter = 10000, maxTime = 15000, maxFunEvals=1e7)
p.fEnough = 1.5e1
p.fTol = 1.0e1
#p.constraints = (y > 5)(tol=1e-4) #x>1e-1 #[2*y<sin(arange(N))]
#r = p.solve(solver, iprint=10, xtol = 1e-36, ftol = 1e-16, show = solver == solvers[-1])
r = p.solve(solver, iprint=10, xtol = 1e-16, ftol = 1e-6, show = solver == solvers[-1])
df = lambda x: [sign(x[0]), sign(x[1])*K, sign(x[2]) * K**2]
#p.df = lambda x: 2*x
#p.plot = 0
#p.xlim = (inf, 5)
#p.ylim = (0, 5000000)
#p.checkdf()
solvers = ['r2', 'ipopt', 'algencan','ralg']
solvers = ['r2', 'algencan','ralg']
#solvers = ['ralg', 'r2']
solvers = ['r2', 'lincher']
solvers = ['ralg']
solvers = ['r2']
#solvers = ['scipy_slsqp']
#solvers = ['algencan']
#solvers = ['ipopt']
colors = ['r', 'b', 'k', 'g']
maxIter = 1000
for i, solver in enumerate(solvers):
p = NSP(f, x0, df=df, xtol = 1e-11, ftol=1e-10, maxIter = maxIter, maxTime=150)
#p.checkdf()
r = p.solve(solver, maxVectorNum=4, iprint=1, showLS=0, plot=0, color=colors[i], show=solver==solvers[-1]) # ralg is name of a solver
#for i, solver in enumerate(solvers):
# p2 = NSP(f, r.xf, df=df, xtol = 1e-6, maxIter = 1200, maxTime=150, ftol=1e-6)
# #p.checkdf()
# r2 = p2.solve(solver, maxVectorNum=15, iprint=1, showLS=1, plot=0, color=colors[i], show=solver==solvers[-1]) # ralg is name of a solver
#print 'x_opt:\n', r.xf
print 'f_opt:', r.ff # should print small positive number like 0.00056
cb.stat['dist'].append(tmp)
cb.stat['f'].append(p.nEvals['f'])
cb.stat['df'].append(-p.nEvals['df'])
return False
asa = lambda x: asarray(x).reshape(-1, 1)
Colors = ['r', 'k', 'b']
lines = []
R = {}
for i, solver in enumerate(solvers):
p = NSP(obj,
startPoint,
fixedVars=(x[0], x[-1]),
maxTime=20,
name='Rzhevsky7 (nVars: ' + str(n) + ')',
maxFunEvals=1e7,
color=Colors[i])
p._prepare()
p.c = None
#p.fEnough = 2.08983385058799+4e-10
p.fOpt = obj(T_optPoint)
p.fTol = 0.5e-15
cb.TMP = 1000
cb.stat = {'dist': [], 'f': [], 'df': []}
r = p.solve(solver,
iprint=10,
xtol=1e-10,
ftol=1e-16,
gtol=1e-10,
