def ex3test():
funcstrdict= {"u2":"u1* (r2+r3)/(r1+r2+r3)", "u3":"u1* r3/(r1+r2+r3)"}
xvectorlistsdict = {"u1":[100], "r1":[20], "r2":[30], "r3":[400]}
spreadvarslist = ["r1", "r2", "r3"]
V=np.array ( [[4, 2, 3],
[2, 9, 6],
[3, 6, 16]])
resdict=sg.generate (funcstrdict, xvectorlistsdict, spreadvarslist, V, 10000, nvoly=1)
ksu2=0.955
ksu3=0.888
diap=0.001
leftksu2=ksu2-diap
rightksu2=ksu2+diap
leftksu3=ksu3-diap
rightksu3=ksu3+diap
print (makepercent(leftksu2, rightksu2, leftksu3, rightksu3, resdict, draw=1))
#first - y axiss, second - x axis
#xvectorlistsdictc = {"u1":100, "r1":20, "r2":30, "r3":400}
#print ("center of the ellise\n",eval(funcstrdict["u2"], xvectorlistsdictc), "\n", eval(funcstrdict["u3"], xvectorlistsdictc) )
def mntkrleq():
#в предположении,что имеет смысл держать центр эллипса совмещённым с центром прямоугольника
global funcstrdict
global spreadvarslist, V
ksu2=0.955
ksu3=0.888
diap=0.001
leftksu2=ksu2-diap
rightksu2=ksu2+diap
leftksu3=ksu3-diap
rightksu3=ksu3+diap
bestpercent=0
vectr=[]
for r2 in range(1,500, 1):
r3=ksu3*r2/(ksu2-ksu3)
r1=r2/(ksu2-ksu3)-r2-r3
xvectorlistsdict = {"u1":[100], "r1":[r1], "r2":[r2], "r3":[r3]}
xvectorlistsdictc = {"u1":100, "r1":r1, "r2":r2, "r3":r3}
#для проверки
centerksu2=eval(funcstrdict["u2"], xvectorlistsdictc)
centerksu3=eval(funcstrdict["u3"], xvectorlistsdictc)
print (r1, r2, r3)
resdict=sg.generate (funcstrdict, xvectorlistsdict, spreadvarslist, V, 1000, nvoly=1)
mkprc=ex3.makepercent(leftksu2, rightksu2, leftksu3, rightksu3, resdict, draw=0)
print ("Percent of good products: ", mkprc[0])
if max(bestpercent, mkprc[0])!=bestpercent:
bestpercent=max (bestpercent, mkprc[0])
vectr=r1,r2,r3
print (bestpercent, "\n", vectr)
#отображаем лучший вариант
xvectorlistsdict = {"u1":[100], "r1":[r1], "r2":[r2], "r3":[r3]}
resdict=sg.generate (funcstrdict, xvectorlistsdict, spreadvarslist, V, 10000, nvoly=1)
mkprc=ex3.makepercent(leftksu2, rightksu2, leftksu3, rightksu3, resdict, draw=1)
def mntkrl():
#Monte-Karlo method V LOB:
bestpercent=0
vectr=[]
for r1 in range(10,500, 1):
for r2 in range(10,500, 1):
for r3 in range(10,500, 1):
funcstrdict= {"u2":"u1* (r2+r3)/(r1+r2+r3)", "u3":"u1* r3/(r1+r2+r3)"}
xvectorlistsdict = {"u1":[100], "r1":[r1], "r2":[r2], "r3":[r3]}
xvectorlistsdictc = {"u1":100, "r1":r1, "r2":r2, "r3":r3}
spreadvarslist = ["r1", "r2", "r3"]
V=np.array ( [[4, 2, 3],
[2, 9, 6],
[3, 6, 16]])
leftksu2, rightksu2, leftksu3, rightksu3 = 0.88, 0.90, 0.95, 0.96
centerksu2=eval(funcstrdict["u2"], xvectorlistsdictc)
centerksu3=eval(funcstrdict["u3"], xvectorlistsdictc)
centerksu2req=leftksu2+(rightksu2-leftksu2)
centerksu3req=leftksu3+(rightksu3-leftksu3)
border=10
if centerksu2req-border<centerksu2<centerksu2req+border and centerksu3req-border<centerksu3<centerksu3req+border:
pass
else:
continue
print (centerksu2, centerksu3)
resdict=sg.generate (funcstrdict, xvectorlistsdict, spreadvarslist, V, 1000, nvoly=1)
mkprc=makepercent(leftksu2, rightksu2, leftksu3, rightksu3, resdict, draw=0)
print ("Percent of good products: ", mkprc[0])
if max(bestpercent, mkprc[0])!=bestpercent:
bestpercent=max (bestpercent, mkprc[0])
vectr=r1,r2,r3
print ("bestpercent=",bestpercent)
print ("vectr=",vectr)
"""
#test area
funcstrdict = {"ksu2": "(r2)/(r1+r2)"}
xvectorlistsdict = {"r1": [20], "r2": [30]}
spreadvarslist = ["r1", "r2"]
V1 = np.array([[4, 2], [3, 6]])
V = np.array([[4, 0], [0, 6]])
resdict = sg.generate(funcstrdict, xvectorlistsdict, spreadvarslist, V, 10000)
diaps = {"ksu2": [0.6, 0.61]}
filteredresdict = filterseq(resdict, diaps)[0]
seqr1 = showonlyoutputseq(resdict, "r1")
seqr2 = showonlyoutputseq(resdict, "r2")
seqr1f = showonlyoutputseq(filteredresdict, "r1")
seqr2f = showonlyoutputseq(filteredresdict, "r2")
getellipse2D(filteredresdict, "r1", "r2")
exit(0)
plt.figure(1) # Here's the part I need, but numbering starts at 1!
def mntkrleq():
#в предположении,что имеет смысл держать центр эллипса совмещённым с центром прямоугольника
global funcstrdict
global spreadvarslist, V
ksu2 = 0.955
ksu3 = 0.888
diap = 0.001
leftksu2 = ksu2 - diap
rightksu2 = ksu2 + diap
leftksu3 = ksu3 - diap
rightksu3 = ksu3 + diap
bestpercent = 0
vectr = []
for r2 in range(1, 500, 1):
r3 = ksu3 * r2 / (ksu2 - ksu3)
r1 = r2 / (ksu2 - ksu3) - r2 - r3
xvectorlistsdict = {"u1": [100], "r1": [r1], "r2": [r2], "r3": [r3]}
xvectorlistsdictc = {"u1": 100, "r1": r1, "r2": r2, "r3": r3}
#для проверки
centerksu2 = eval(funcstrdict["u2"], xvectorlistsdictc)
centerksu3 = eval(funcstrdict["u3"], xvectorlistsdictc)
print(r1, r2, r3)
resdict = sg.generate(funcstrdict,
xvectorlistsdict,
spreadvarslist,
V,
1000,
nvoly=1)
mkprc = ex3.makepercent(leftksu2,
rightksu2,
leftksu3,
rightksu3,
resdict,
draw=0)
print("Percent of good products: ", mkprc[0])
if max(bestpercent, mkprc[0]) != bestpercent:
bestpercent = max(bestpercent, mkprc[0])
vectr = r1, r2, r3
print(bestpercent, "\n", vectr)
#отображаем лучший вариант
xvectorlistsdict = {"u1": [100], "r1": [r1], "r2": [r2], "r3": [r3]}
resdict = sg.generate(funcstrdict,
xvectorlistsdict,
spreadvarslist,
V,
10000,
nvoly=1)
mkprc = ex3.makepercent(leftksu2,
rightksu2,
leftksu3,
rightksu3,
resdict,
draw=1)
def mntkrleq():
#в предположении,что имеет смысл держать центр эллипса совмещённым с центром прямоугольника
global funcstrdict
global spreadvarslist, V
ksu2=0.955
ksu3=0.888
diap=0.001
leftksu2=ksu2-diap
rightksu2=ksu2+diap
leftksu3=ksu3-diap
rightksu3=ksu3+diap
bestpercent=0
vectr=[]
fig = plt.figure()
plt.xlabel('seqksu2')
plt.ylabel('seqksu3')
plt.ion()
rlist=list() #список коэффициентов корреляции
perclist=list() #список значений процента годности
endrange=100
for r2 in range(1,endrange, 1):
r3=ksu3*r2/(ksu2-ksu3)
r1=r2/(ksu2-ksu3)-r2-r3
xvectorlistsdict = {"u1":[100], "r1":[r1], "r2":[r2], "r3":[r3]}
xvectorlistsdictc = {"u1":100, "r1":r1, "r2":r2, "r3":r3}
#для проверки
centerksu2=eval(funcstrdict["u2"], xvectorlistsdictc)
centerksu3=eval(funcstrdict["u3"], xvectorlistsdictc)
print (r1, r2, r3)
resdict=sg.generate (funcstrdict, xvectorlistsdict, spreadvarslist, V, 1000, nvoly=1)
mkprc=makepercent(leftksu2, rightksu2, leftksu3, rightksu3, resdict, draw=0)
print ("Percent of good products: ", mkprc[0])
perclist.append(mkprc[0])
rlist.append(mkprc[1][1][0])
if max(bestpercent, mkprc[0])!=bestpercent:
bestpercent=max (bestpercent, mkprc[0])
vectr=r1,r2,r3
plt.clf()
# plt.axis([0.94, 0.97, 0.86, 0.92])
plt.plot(mkprc[2], mkprc[3], 'ro') # Returns a tuple of line objects, thus the comma
lplots=[plt.plot ([leftksu2, leftksu2], [leftksu3, rightksu3]),
plt.plot ([rightksu2, rightksu2], [leftksu3, rightksu3]),
plt.plot ([leftksu2, rightksu2], [leftksu3, leftksu3]),
plt.plot ([leftksu2, rightksu2], [rightksu3, rightksu3])]
for p in lplots:
plt.setp(p, color='b', linewidth=2.0)
plt.draw()
time.sleep(0.1)
print (bestpercent, "\n", vectr)
#fig1 = plt.figure(2) #лучший результат
fig2 = plt.figure(2) #график r
plt.subplot(211)
plt.plot(range(1,endrange, 1), rlist)
plt.ylabel("r")
plt.subplot(212)
plt.plot(range(1,endrange, 1), perclist)
plt.ylabel("percent")
plt.show(block=True)
plt.ion()
fig = plt.figure()
if 0:
funcstrdict= {"u2":"u1* (r2+r3)/(r1+r2+r3)", "u3":"u1* r3/(r1+r2+r3)"}
#xvectorlistsdict = {"u1":[100], "r1":[20], "r2":[r2], "r3":[400]}
xvectorlistsdict = {"u1":[100], "r1":[1538], "r2":[920], "r3":[12193]}
spreadvarslist = ["r1", "r2", "r3"]
V=np.array ( [[4, 2, 3],
[2, 9, 6],
[3, 6, 16]])
resdict=sg.generate (funcstrdict, xvectorlistsdict, spreadvarslist, V, 10000, nvoly=1)
leftksu2, rightksu2, leftksu3, rightksu3 = 0.88, 0.90, 0.95, 0.96
mkprc=makepercent(leftksu2, rightksu2, leftksu3, rightksu3, resdict, draw=1)
#print ("r2=",r2)
print ("Percent of good products: ", mkprc[0])
print ("CorrCoef: ", mkprc[1][0][1])
# if mkprc[1][0][1]<0:
# break
plt.xlabel('seqksu2')
plt.ylabel('seqksu3')
#plt.axis([0.94, 0.97, 0.86, 0.92])
def mntkrleq():
#в предположении,что имеет смысл держать центр эллипса совмещённым с центром прямоугольника
global funcstrdict
global spreadvarslist, V
ksu2 = 0.955
ksu3 = 0.888
diap = 0.001
leftksu2 = ksu2 - diap
rightksu2 = ksu2 + diap
leftksu3 = ksu3 - diap
rightksu3 = ksu3 + diap
bestpercent = 0
vectr = []
fig = plt.figure()
plt.xlabel('seqksu2')
plt.ylabel('seqksu3')
plt.ion()
rlist = list() #список коэффициентов корреляции
perclist = list() #список значений процента годности
endrange = 100
for r2 in range(1, endrange, 1):
r3 = ksu3 * r2 / (ksu2 - ksu3)
r1 = r2 / (ksu2 - ksu3) - r2 - r3
xvectorlistsdict = {"u1": [100], "r1": [r1], "r2": [r2], "r3": [r3]}
xvectorlistsdictc = {"u1": 100, "r1": r1, "r2": r2, "r3": r3}
#для проверки
centerksu2 = eval(funcstrdict["u2"], xvectorlistsdictc)
centerksu3 = eval(funcstrdict["u3"], xvectorlistsdictc)
print(r1, r2, r3)
resdict = sg.generate(funcstrdict,
xvectorlistsdict,
spreadvarslist,
V,
1000,
nvoly=1)
mkprc = makepercent(leftksu2,
rightksu2,
leftksu3,
rightksu3,
resdict,
draw=0)
print("Percent of good products: ", mkprc[0])
perclist.append(mkprc[0])
rlist.append(mkprc[1][1][0])
if max(bestpercent, mkprc[0]) != bestpercent:
bestpercent = max(bestpercent, mkprc[0])
vectr = r1, r2, r3
plt.clf()
# plt.axis([0.94, 0.97, 0.86, 0.92])
plt.plot(mkprc[2], mkprc[3],
'ro') # Returns a tuple of line objects, thus the comma
lplots = [
plt.plot([leftksu2, leftksu2], [leftksu3, rightksu3]),
plt.plot([rightksu2, rightksu2], [leftksu3, rightksu3]),
plt.plot([leftksu2, rightksu2], [leftksu3, leftksu3]),
plt.plot([leftksu2, rightksu2], [rightksu3, rightksu3])
]
for p in lplots:
plt.setp(p, color='b', linewidth=2.0)
plt.draw()
time.sleep(0.1)
print(bestpercent, "\n", vectr)
#fig1 = plt.figure(2) #лучший результат
fig2 = plt.figure(2) #график r
plt.subplot(211)
plt.plot(range(1, endrange, 1), rlist)
plt.ylabel("r")
plt.subplot(212)
plt.plot(range(1, endrange, 1), perclist)
plt.ylabel("percent")
plt.show(block=True)
def mntkrl():
#Monte-Karlo method V LOB:
bestpercent = 0
vectr = []
for r1 in range(10, 500, 1):
for r2 in range(10, 500, 1):
for r3 in range(10, 500, 1):
funcstrdict = {
"u2": "u1* (r2+r3)/(r1+r2+r3)",
"u3": "u1* r3/(r1+r2+r3)"
}
xvectorlistsdict = {
"u1": [100],
"r1": [r1],
"r2": [r2],
"r3": [r3]
}
xvectorlistsdictc = {"u1": 100, "r1": r1, "r2": r2, "r3": r3}
spreadvarslist = ["r1", "r2", "r3"]
V = np.array([[4, 2, 3], [2, 9, 6], [3, 6, 16]])
leftksu2, rightksu2, leftksu3, rightksu3 = 0.88, 0.90, 0.95, 0.96
centerksu2 = eval(funcstrdict["u2"], xvectorlistsdictc)
centerksu3 = eval(funcstrdict["u3"], xvectorlistsdictc)
centerksu2req = leftksu2 + (rightksu2 - leftksu2)
centerksu3req = leftksu3 + (rightksu3 - leftksu3)
border = 10
if centerksu2req - border < centerksu2 < centerksu2req + border and centerksu3req - border < centerksu3 < centerksu3req + border:
pass
else:
continue
print(centerksu2, centerksu3)
resdict = sg.generate(funcstrdict,
xvectorlistsdict,
spreadvarslist,
V,
1000,
nvoly=1)
mkprc = makepercent(leftksu2,
rightksu2,
leftksu3,
rightksu3,
resdict,
draw=0)
print("Percent of good products: ", mkprc[0])
if max(bestpercent, mkprc[0]) != bestpercent:
bestpercent = max(bestpercent, mkprc[0])
vectr = r1, r2, r3
print("bestpercent=", bestpercent)
print("vectr=", vectr)
#test area
import sequence_generation as sg
#funcstrdict= {"y1":"u1* (r2+r3)/(r1+r2+r3)", "y2":"u1* r3/(r1+r2+r3)"}
funcstrdict= {"y1":"u1* (r2+r3)", "y2":"u1* r3"}
#xvectorlistsdict = {"u1":range(1,10), "r1":[20], "r2":[30], "r3":[400]}
xvectorlistsdict = {"u1":range(1,10), "r2":[30], "r3":[400]}
vrslst=sg.generate (funcstrdict, xvectorlistsdict, None, Vx=None, nvolx=None, yvectordispsdict=None, nvoly=1, outfilename="t.txt", listoutvars=["y1", "y2", "u1"] )
#import pickle
#pickle.dump(vrslst, open("vrslt.f", "wb"))
#vrslst=pickle.load(open("vrslt.f", "rb"))
#сюда впилить чтение файла
grandCountGN(funcstrdict,["u1"] , ["y1", "y2"],["r2", "r3"] , vrslst, NSIG=5, kinit=np.array( [ 1, 2]))
#grandCountGN(funcstrdict,["u1"] , ["y1", "y2"],["r1", "r2", "r3"] , vrslst, NSIG=5, kinit=np.array( [21, 31, 401]))
##grandCountGN(funcstrdict,["u1"] , ["y1", "y2"],["r1", "r23"] , vrslst, NSIG=5)
