-
Notifications
You must be signed in to change notification settings - Fork 0
/
pyDONEc.py
90 lines (74 loc) · 3.86 KB
/
pyDONEc.py
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
import numpy
import ctypes
import os
class optimizer():
#initialization, parameters are:
#startvect: Initial guess of the minimum position
#d=degrees of freedom
#lb=global lower bound, or ndarray of lower bounds (the length must be equal to the number of degrees of freedom)
#ub=global upper bound, or ndarray of upper bounds (the length must be equal to the number of degrees of freedom)
#D=number of basis functions
#lam=regularization parameter
#sigma=expected width of the metric function spectra
#expl=size of the exploration steps
#memory= size of the memory buffer for the function fit
def __init__(self,startvect,d,lb,ub,D,lam,sigma,expl,memory):
#load the done step function from the dll
cwd = os.getcwd()
os.chdir("dlls")
DONEdll = ctypes.cdll.LoadLibrary("DONEdll.dll")
self.DONEvect= getattr(DONEdll,"?DONE@@YAHPEAN0H00NNN0HH@Z")
os.chdir(cwd)
#store values of the parameters
self.d=int(d)
self.D=ctypes.c_double(D)
self.sigma=ctypes.c_double(sigma)
self.lam=ctypes.c_double(lam)
self.memory=int(memory)
if type(lb)==numpy.ndarray:
self.lb=lb
else:
self.lb=numpy.ones(d)*lb
if type(ub)==numpy.ndarray:
self.ub=ub
else:
self.ub=numpy.ones(d)*ub
if type(expl)==numpy.ndarray:
self.expl=expl
else:
self.expl=numpy.ones(d)*expl
self.reset=True
#converts Python ndarrays to pointers to c++ vectors
self.lbpoint=self.lb.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
self.ubpoint=self.ub.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
self.explpoint=self.expl.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
#creates the expected minimum coordinates and next step information, and converts in c++ vectors
self.x0=startvect
self.x0point=self.x0.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
#the next step information has size d+1, and contains the parameters used for the measurement, followed by the corresponding value of the metric
self.measinfo=numpy.zeros(self.d+1)
self.measinfo[0:self.d]=self.x0
self.measinfopoint=self.measinfo.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
#step function of the algorithm, evalulates the passed function in the next step location and updates x0 and measinfo
def step(self,function):
self.measinfo[self.d]=function(self.measinfo[0:self.d])
if self.reset:
self.DONEvect(self.measinfopoint,self.x0point,self.d,self.lbpoint,self.ubpoint,self.D,self.lam,self.sigma,self.explpoint,self.memory,1)
self.reset=False
else:
self.DONEvect(self.measinfopoint,self.x0point,self.d,self.lbpoint,self.ubpoint,self.D,self.lam,self.sigma,self.explpoint,self.memory,0)
#null step function of the algorithm, performs a step in a user defined location, with a user defined corresponding value of the metric function
def nullstep(self,x,metric):
self.measinfo[self.d]=metric
self.measinfo[0:self.d]=x
if self.reset:
self.DONEvect(self.measinfopoint,self.x0point,self.d,self.lbpoint,self.ubpoint,self.D,self.lam,self.sigma,self.explpoint,self.memory,1)
self.reset=False
else:
self.DONEvect(self.measinfopoint,self.x0point,self.d,self.lbpoint,self.ubpoint,self.D,self.lam,self.sigma,self.explpoint,self.memory,0)
#returns the expected position of the minimum
def getmin(self):
return self.x0
#returns the most recent dataset and metric value
def getlaststep(self):
return self.measinfo[0:self.d],self.measinfo[self.d]