我正在尝试开发一种算法来解决cplex中的特定混合整数问题。
当我尝试运行算法时，得到以下文本：
Traceback (most recent call last):
Here the log shows the directory of the files of my algorithm for a few lines
.
.
.
.
and finally:
File "/anaconda2/lib/python2.7/site-packages/cplex/__init__.py", line 1099, in solve
_proc.mipopt(self._env._e, self._lp)
File "/anaconda2/lib/python2.7/site-packages/cplex/_internal/_procedural.py", line 629, in mipopt
check_status(env, status)
File "/anaconda2/lib/python2.7/site-packages/cplex/_internal/_procedural.py", line 303, in __call__
raise CplexSolverError(error_string, env, status)
cplex.exceptions.errors.CplexSolverError: CPLEX Error 1006: Error during callback.
您是否知道可能导致该文件错误的原因？
我有一些代码文件：
这是启动分辨率的文件，称为G_resoluciones：
import G_Callbacks
import G_Heuristica
import gc
import D2MIPpG
import cplex
import os
def ResolvemosBC_q_C_norm(E,I,J,K,p,R,C):
D2_C_DrBC_C_norm = open(os.path.abspath(os.path.join(os.getcwd(), '..'))+"/G_ResultsD2_C_DrBC_Cp_norm.txt", "a")
tiemposD2_C_DrBC_norm = {}
tiemposRelD2_C_DrBC_norm = {}
nodosD2_C_DrBC_norm = {}
CODD2_C_DrBC_norm={}
ValObjD2_C_DrBC_norm = {}
ValRelD2_C_DrBC_norm = {}
MSJD2_C_DrBC_norm = {}
MIPGAPD2_C_DrBC_norm = {}
OptimalityGapD2_C_DrBC_norm = {}
CutsD2_C_DrBC_norm = {}
IterD2_C_DrBC_norm = {}
BestBoundD2_C_DrBC_norm = {}
IncD2_C_DrBC_norm = {}
LP_GORDO_C_DrBC_norm = {}
sH,thetaH,xH,qH,fobjH={},{},{},{},{}
m1 = D2MIPpG.solveD2(I,J,K,p,R,C)
sH,thetaH,xH,qH,fobjH=G_Heuristica.HeuristicaGeneral(p,I,K,J,R,C)
qHH={}
indi=[]
for k in range(K):
for j in range(J):
qHH[J*k+j]=qH[k,j]
coefi=[]
for i in range(I):
coefi.append(xH[i])
for j in range(I,I+J*K,1):
coefi.append(qHH[j-I])
for j in range(I+J*K,I+J*K+K,1):
coefi.append(sH[j-I-J*K])
for j in range(I+J*K+K,I+J*K+K+K,1):
coefi.append(thetaH[j-I-J*K-K])
for i in range(len(coefi)):
indi.append(i)
m1.MIP_starts.add(cplex.SparsePair(ind=indi, val=coefi), m1.MIP_starts.effort_level.solve_fixed,"primera") # primeros valores
Nodos_cb = m1.register_callback(G_Callbacks.Nodos)
Nodos_cb.times_called = 0
Inc=m1.register_callback(G_Callbacks.Incumbente)
G_Callbacks.Cplex_callback_D2BC_C_norm.model=m1
G_Callbacks.Cplex_callback_D2BC_C_norm.thetaVars=m1._data[3]
G_Callbacks.Cplex_callback_D2BC_C_norm.qVars=m1._data[1]
e=m1.register_callback(G_Callbacks.Cplex_callback_D2BC_C_norm)
e.iterations=0
e.iterationsR=0
m1.parameters.mip.cuts.mircut.set(-1)
m1.parameters.mip.cuts.gomory.set(-1)
m1.parameters.mip.limits.treememory.set(50000)
m1.parameters.mip.strategy.file.set(3)
m1.parameters.workmem.set(10000)
m1.parameters.timelimit.set(18000)
m1.parameters.preprocessing.presolve.set(0)
m1.parameters.preprocessing.numpass.set(0)
m1.parameters.threads.set(1)
m1.parameters.mip.strategy.fpheur.set(-1)
m1.parameters.mip.strategy.heuristicfreq.set(-1)
m1.parameters.mip.strategy.lbheur.set(0)
m1.parameters.mip.strategy.rinsheur.set(-1)
# m5._alpha=alpha
ValorObjetivoMejorado={}
Counter={}
m1._Counter = Counter
m1._ValorObjetivoMejorado = ValorObjetivoMejorado
m1._qH=qH
m1._thetaH=thetaH
m1._fobjH=fobjH
m1._K = K
m1._J = J
m1._p = p
m1._I = I
m1._R = R
m1._C = C
m1._FirstTime=1
start = m1.get_time()
m1.solve() #this is where the spyder says the error is
end = m1.get_time()
elapsed = end - start
solrEC_D=m1.solution
elapsedLP=e.stopic-start
ValObjD2_C_DrBC_norm,tiemposD2_C_DrBC_norm, nodosD2_C_DrBC_norm, MSJD2_C_DrBC_norm, CODD2_C_DrBC_norm,ValRelD2_C_DrBC_norm, tiemposRelD2_C_DrBC_norm, MIPGAPD2_C_DrBC_norm,OptimalityGapD2_C_DrBC_norm, CutsD2_C_DrBC_norm,IterD2_C_DrBC_norm,BestBoundD2_C_DrBC_norm,LP_GORDO_C_DrBC_norm=Escribe_Resultados_CUTANDBRANCH(m1,Nodos_cb,e,elapsed,elapsedLP)
D2_C_DrBC_C_norm.write('G_B&C_q_norm')
D2_C_DrBC_C_norm.write(' %d %d %d %d ' % (E, I, J, K))
D2_C_DrBC_C_norm.write(' %0.4f %0.4f %d %s %d %0.4f %0.4f %0.4f' % (float(ValObjD2_C_DrBC_norm), float(tiemposD2_C_DrBC_norm), nodosD2_C_DrBC_norm, MSJD2_C_DrBC_norm,CODD2_C_DrBC_norm, float(ValRelD2_C_DrBC_norm), float(tiemposRelD2_C_DrBC_norm), float(MIPGAPD2_C_DrBC_norm)))
D2_C_DrBC_C_norm.write(' %0.4f %d %d %0.4f' % (float(OptimalityGapD2_C_DrBC_norm), CutsD2_C_DrBC_norm,IterD2_C_DrBC_norm,float(BestBoundD2_C_DrBC_norm)))
D2_C_DrBC_C_norm.write(' %0.4f ' %(LP_GORDO_C_DrBC_norm))
D2_C_DrBC_C_norm.write(' 1')
D2_C_DrBC_C_norm.write("\n")
del m1
gc.collect()
def Escribe_Resultados_CUTANDBRANCH(modelo, callback1, callback2, elapsed, elapsedLP):
ValObj,tiempos,nodos, MSJ,COD,ValRel,tiemposRel,MIPGAP,OptimalityGap,Cuts,Iter,BestBound,LP_GORDO = {},{},{},{},{},{},{},{},{},{},{},{},{}
solrEC_D=modelo.solution
#OPTIMAL INTEGER SOLUTION FOUND OR OPTIMAL SOLUTION WITH EPGAP OR EPAGAP TOLERANCE FOUND
if solrEC_D.get_status()==101 or solrEC_D.get_status()==102:
ValObj=solrEC_D.get_objective_value()
tiempos=elapsed
nodos=callback2.nodecnt
MSJ=solrEC_D.status[solrEC_D.get_status()]
COD=solrEC_D.get_status()
ValRel=callback2.LPValueWithCuts
tiemposRel=elapsedLP
MIPGAP=((ValRel-ValObj)/ValObj)*100
OptimalityGap=0
Cuts=callback2.cuts+1
Iter=callback2.iterationsR
BestBound=callback2.mipnodeobjbnd
LP_GORDO=callback2.TimeBetweenCalls
#TIMELIMIT REACHED BUT INTEGER SOL EXISTS
if solrEC_D.get_status()==107:
ValObj=solrEC_D.get_objective_value()
tiempos=elapsed
nodos=-1
MSJ=solrEC_D.status[solrEC_D.get_status()]
COD=solrEC_D.get_status()
ValRel=callback2.LPValueWithCuts
tiemposRel=elapsedLP
MIPGAP=-1
OptimalityGap=callback1.MIP_relative_gap
Cuts=callback2.cuts+1
Iter=callback2.iterationsR
BestBound=callback2.mipnodeobjbnd
LP_GORDO=callback2.TimeBetweenCalls
#TIMELIMIT REACHED BUT INTEGER SOL DOES NOT EXIST
if solrEC_D.get_status()==108:
ValObj=-1
tiempos=elapsed
nodos=-1
MSJ=solrEC_D.status[solrEC_D.get_status()]
COD=solrEC_D.get_status()
ValRel=callback2.LPValueWithCuts
tiemposRel=elapsedLP
MIPGAP=-1
OptimalityGap=-1
Cuts=callback2.cuts+1
Iter=callback2.iterationsR
BestBound=callback2.mipnodeobjbnd
LP_GORDO=callback2.TimeBetweenCalls
#INFEASIBLE OR UNBOUNDED
if solrEC_D.get_status()==119:
ValObj=-1
tiempos=elapsed
nodos=-1
MSJ=solrEC_D.status[solrEC_D.get_status()]
COD=solrEC_D.get_status()
ValRel=-1
tiemposRel=-1
MIPGAP=-1
OptimalityGap=-1
Cuts=-1
Iter=-1
BestBound=-1
LP_GORDO=0
return ValObj,tiempos,nodos, MSJ,COD,ValRel,tiemposRel,MIPGAP,OptimalityGap,Cuts,Iter,BestBound,LP_GORDO
现在，第一个调用的其他代码文件：
G_callbacks：
from __future__ import division
import cplex.callbacks
from cplex.callbacks import MIPInfoCallback
import G_SeparationProblems
import numpy
import operator
#MIPInfoCallback
class Nodos(MIPInfoCallback):
num_nodes=-1
best_OBJ=-1
MIP_relative_gap=-1
times_called =0
LPSol= -1
FirstT = 1
stopinc=0
#These values are set in case the problem is solved at the first iteration of the root, because if that happens I guess
#we never even get to call the callback. In any case, that never happens for my instances.
def __call__(self):
self.times_called += 1
self.num_nodes=self.get_num_nodes()
self.best_OBJ=self.get_best_objective_value()
self.MIP_relative_gap=self.get_MIP_relative_gap()
if self.times_called==1 and self.FirstT == 1:
self.LPSol=self.get_best_objective_value()
self.stopinc=self.get_time()
self.FirstT = 0
#IncumbentCallback
class Incumbente(cplex.callbacks.IncumbentCallback):
def __call__(self):
print "El callback Incumbente dice: "
print self.get_solution_source()
class Cplex_callback_D2BC_C_norm(cplex.callbacks.UserCutCallback):
def __init__(self,env):
cplex.callbacks.UserCutCallback.__init__(self,env)
self.env=env
LPValueWithCuts=-1
cuts=-1
iterations=0
lb=-cplex.infinity
ub=cplex.infinity
lbPrimal=-cplex.infinity
iterationsR=0
times_called=0
BeenOptimal=0
alpha = 0
q_sep={}
theta_sep={}
q_in={}
theta_in={}
cbasis={}
vbasis={}
te,ue,ve={},{},{}
tr,ur,vr={},{},{}
startic={}
RepDif= 0
Dif=0
DifAnterior=0
HaHabidoCortes=0
TimeBetweenCalls=-1
UES={}
max_index={}
max_value={}
q_puedeser={}
theta_puedeser={}
fobj_puedeser={}
cambio={}
def __call__(self):
if self.model._Counter == 1:
print "EN EL USER CUT CALLBACK : %s" %self.model._ValorObjetivoMejorado
Vars=sorted(self.thetaVars.values())
coeffs = {}
for k in range(self.model._K):
if not self.thetaVars[k] in coeffs:
coeffs[self.thetaVars[k]]=0
coeffs[self.thetaVars[k]]+=self.model._p[k]
self.add(cut=cplex.SparsePair(ind=coeffs.keys(),val=coeffs.values()),sense="L", rhs=self.model._ValorObjetivoMejorado,use=0)
self.model._Counter = 0
self.ub=self.get_objective_value()
print "ITERACION %s" %(self.times_called)
self.nodecnt=self.get_num_nodes()
print "Callback called %s times" %self.times_called
if self.times_called==0:
for k in range(self.model._K):
for j in range(self.model._J):
self.q_in[k,j]=self.model._qH[k,j]
for k in range(self.model._K):
self.theta_in[k]=self.model._thetaH[k]
self.DifAnterior=self.ub-self.lb
if self.RepDif >= 30:
print "SALTANDO QUE ES GERUNDIO!"
self.lb=self.ub
self.times_called+=1
self.startic[self.times_called] = self.model.get_time()
if self.nodecnt == 0 and self.times_called>=2:
self.TimeBetweenCalls=self.startic[self.times_called]-self.startic[self.times_called-1]
r=1
self.mipnodeobjbnd = self.get_objective_value()
if self.nodecnt == 1 and self.model._FirstTime == 1:
self.stopic = self.model.get_time()
self.model._FirstTime=0
if self.nodecnt >= 0:
self.stopic = self.model.get_time()
Vars2=sorted(self.thetaVars.values())
theta=self.get_values(Vars2)
Vars=sorted(self.qVars.values())
a=self.get_values(Vars)
data=numpy.array(a)
shape=(self.model._K, self.model._J)
q=data.reshape(shape)
sub={}
self.lbPrimal=self.get_incumbent_objective_value()
tr,ur,vr={},{},{}
self.CurrentBestObjBound=self.get_best_objective_value()
self.CurrentBestObj=self.get_incumbent_objective_value()
if self.ub-self.lbPrimal > 1e-4*self.ub:
self.alpha = 1
for k in range(self.model._K):
for j in range(self.model._J):
self.q_sep[k,j]=self.alpha*q[k,j]+(1-self.alpha)*self.q_in[k,j]
for k in range(self.model._K):
self.theta_sep[k]=self.alpha*theta[k]+(1-self.alpha)*self.theta_in[k]
print self.ub
print self.lb
if self.ub-self.lb > 1e-4*self.ub:
self.HaHabidoCortes=0
for k in range(self.model._K):
h=0
sub=G_SeparationProblems.SeparationProblem_q_C_norm(self.model._p,self.q_sep,self.theta_sep,self.model._I,self.model._J,self.model._K,self.model._R,self.model._C)
sub.set_problem_type(sub.problem_type.LP)
print "El Problema es un %s: %s" %(sub.get_problem_type(),sub.problem_type[sub.get_problem_type()])
sub.parameters.preprocessing.presolve.set(sub.parameters.preprocessing.presolve.values.off)
sub.parameters.lpmethod.set(sub.parameters.lpmethod.values.primal)
sub.solve()
if sub.solution.get_status()==sub.solution.status.optimal:
print "Estamos dentro para recuperar vertice"
a,aa={},{}
verticealpha={}
a=sub.solution.get_values(range(0,self.model._K*self.model._I,1))
data=numpy.array(a)
shape=(self.model._K,self.model._I)
aa=data.reshape(shape)
for k in range(self.model._K):
for i in range(self.model._I):
verticealpha[k,g,i]=aa[k,i]
b,bb={},{}
verticebeta={}
b=sub.solution.get_values(range(self.model._K*self.model._I, self.model._K*self.model._I+self.model._K*self.model._J,1))
data=numpy.array(b)
shape=(self.model._K,self.model._J)
bb=data.reshape(shape)
for k in range(self.model._K):
for j in range(self.model._J):
verticebeta[k,g,j]=bb[k,j]
c,cc={},{}
verticegamma={}
c=sub.solution.get_values(range(self.model._K*self.model._I+self.model._K*self.model._J, self.model._K*self.model._I+self.model._K*self.model._J+self.model._K*self.model._J*self.model._J,1))
data=numpy.array(c)
shape=(self.model._K, self.model._J, self.model._J)
cc=data.reshape(shape)
for k in range (self.model._K):
for l in range (self.model._J):
for j in range (self.model._J):
verticegamma[k,g,l,j]=cc[k,l,j]
ll={}
verticelamb={}
ll=sub.solution.get_values(range(self.model._K*self.model._I+self.model._K*self.model._J+self.model._K*self.model._J*self.model._J,self.model._K*self.model._I+self.model._K*self.model._J+self.model._K*self.model._J*self.model._J+self.model._K,1))
for k in range(self.model._K):
verticelamb[k,g]=ll[k]
h=h+1
self.iterations=self.iterations+1
print "h=%s" %h
if h!=0:
for a in range(h-1,h,1):
if(sum(self.theta_sep[k]*(1+self.model._p[k]*verticelamb[k,a]) for k in range(self.model._K))-1e-4> sum(self.q_sep[k,j]*verticebeta[k,a,j] for k in range(self.model._K) for j in range(self.model._J))):
self.HaHabidoCortes=1
r=r+1
coeffs={}
for k in range(self.model._K):
if not self.thetaVars[k] in coeffs:
coeffs[self.thetaVars[k]]=0
coeffs[self.thetaVars[k]]+=(1+self.model._p[k]*verticelamb[k,a])
for k in range(self.model._K):
for j in range(self.model._J):
if not self.qVars[k,j] in coeffs:
coeffs[self.qVars[k,j]]=0
coeffs[self.qVars[k,j]]-=verticebeta[k,a,j] #0 mas eso
self.add(cut=cplex.SparsePair(ind=coeffs.keys(),val=coeffs.values()), sense="L", rhs=0, use=self.use_cut.force)
print "CORTE DE OPTIMALIDAD AÑADIDO"
self.iterationsR += 1
self.lb=sub.solution.get_objective_value()
print "HAHABIDOCORTES: %s" %self.HaHabidoCortes
if self.HaHabidoCortes==0:
for k in range(self.model._K):
for j in range(self.model._J):
self.q_in[k,j]=self.q_sep[k,j]
for k in range(self.model._K):
self.theta_in[k]=self.theta_sep[k]
self.Dif=self.ub-self.lb
if self.DifAnterior-self.Dif==0:
self.RepDif+=1
print "DifAnterior:= %s -- Dif:= %s" %(self.DifAnterior, self.Dif)
else:
self.lb = -cplex.infinity
print "**************************"
print "NODE: %s" %self.nodecnt
print "UPPER BOUND: %s" %self.ub
print "LOWER BOUND D: %s" %self.lb
print "LOWER BOUND P: %s" %self.lbPrimal
print "**************************"
self.cuts=self.get_num_cuts(134) #134 is user cuts
self.startic[self.times_called] = self.model.get_time()
if self.nodecnt>=0:
self.LPValueWithCuts=sum(self.model._p[k]*theta[k] for k in range(self.model._K))
self.model._LPValueWithCuts=self.LPValueWithCuts
G_heuristica
import operator
import RANDUtilidades
import random
def HeuristicaGeneral(p,I,K,J,R,C):
x={}
q={}
UES={}
max_index={}
max_value={}
d={}
s={}
fobj={}
for i in range(I):
x[i]=float(1)
for i in range(I):
x[i]=x[i]/I
for k in range(K):
for j in range(J):
UES[k,j]=sum(C[k,i,j]*x[i] for i in range(I))
for k in range(K):
max_index[k], max_value[k] = max(enumerate(UES[k,j] for j in range(J)), key=operator.itemgetter(1))
for k in range(K):
for j in range(J):
if j == max_index[k]:
q[k,j]=1
else:
q[k,j]=0
for k in range(K):
d[k]=sum(R[k,i,max_index[k]]*x[i] for i in range(I))
for k in range(K):
s[k]=sum(C[k,i,max_index[k]]*x[i] for i in range(I))
fobj=sum(p[k]*d[k] for k in range(K))
return s,d,x,q,fobj
D2MIPG：
import cplex
def solveMIPpG(I,J,K,p,R,C):
model = cplex.Cplex()
#Definition of the variables
x, q, z = {},{},{}
for i in range(I):
x[i]=model.variables.add(obj=[0],lb=[0],ub=[1],types=["C"], names=["x[%s]" %i])[0]
for k in range(K):
for j in range(J):
q[k,j]=model.variables.add(obj=[0],lb=[0], types=["B"], names=["q[%s,%s]" %(k,j)])[0]
for k in range(K):
for j in range(J):
for i in range(I):
z[k,i,j]=model.variables.add(obj=[p[k]*R[k,i,j]], lb=[0], types=["C"], names=["z[%s,%s,%s]" %(k,i,j)])[0]
model.objective.set_sense(model.objective.sense.maximize)
#Model constraints
#Suma qjs igual a 1.
for k in range(K):
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=[q[k,j] for j in range(J)],val=[1.0]*J)], senses=["E"], rhs=[1], names=['SumaProbs[%s]' %k])[0]
for k in range(K):
for j in range(J):
coef=[]
for i in range(I):
coef.append(1)
coef.append(-1)
var=[]
for i in range(I):
var.append(z[k,i,j])
var.append(q[k,j])
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=var, val=coef)], senses=["E"], rhs=[0], names=['sumazsq[%s,%s]'%(k,j)])[0]
for k in range(K):
for i in range(I):
coef2=[]
for j in range(J):
coef2.append(1)
coef2.append(-1)
var2=[]
for j in range(J):
var2.append(z[k,i,j])
var2.append(x[i])
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=var2, val=coef2)], senses=["E"], rhs=[0], names=['sumazsx[%s,%s]'%(k,i)])[0]
for k in range(K):
for j in range(J):
for l in range(J):
if j!=l:
coef3=[]
for i in range(I):
coef3.append(C[k,i,l]-C[k,i,j])
var3=[]
for i in range(I):
var3.append(z[k,i,j])
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=var3, val=coef3)], senses=["L"], rhs=[0], names=['bestresp[%s,%s,%s]'%(k,l,j)])[0]
model._data = x, q, z
# model.write("MIPpGHAHA.lp")
return model
#This is the weak formulation D2 #Generador inicial del modelo
def solveD2(I,J,K,p,R,C):
model = cplex.Cplex()
#Definition of the variables
theta, s ,q, x = {},{},{},{}
for i in range(I):
x[i]=model.variables.add(obj=[0],lb=[0],ub=[1],types=["C"], names=["x[%s]" %i])[0]
for k in range(K):
for j in range(J):
q[k,j]=model.variables.add(obj=[0],lb=[0], types=["B"], names=["q[%s,%s]" %(k,j)])[0]
for k in range(K):
s[k] = model.variables.add(obj=[0], types=["C"], names=["s[%s]" % k])[0]
for k in range(K):
theta[k] = model.variables.add(obj=[p[k]], types=["C"], names=["theta[%s]" % k])[0]
model.objective.set_sense(model.objective.sense.maximize)
M1, M2= {}, {}
M1 = 2 * max(R.values()) + 1
M2 = 2 * max(C.values()) + 1
#Constraints of the model
#Suma qjs igual a 1.
for k in range(K):
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=[q[k,j] for j in range(J)],val=[1.0]*J)], senses=["E"], rhs=[1], names=['SumaProbs[%s]' %k])[0]
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=[x[i] for i in range(I)],val=[1.0]*I)], senses=["E"], rhs=[1], names=['Sumaxs'])[0]
for k in range(K):
for j in range(J):
coef=[]
coef.append(1)
for i in range(I):
coef.append(-R[k,i,j])
coef.append(M1)
vars=[]
vars.append(theta[k])
for i in range(I):
vars.append(x[i])
vars.append(q[k,j])
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=vars, val=coef)], senses=["L"], rhs=[M1], names=['Bound_d[%s,%s]' %(k,j)])[0]
for k in range(K):
for j in range(J):
coef2=[]
coef2.append(1)
for i in range(I):
coef2.append(-C[k,i,j])
coef2.append(M2)
vars2=[]
vars2.append(s[k])
for i in range(I):
vars2.append(x[i])
vars2.append(q[k,j])
coef3=[]
coef3.append(1)
for i in range(I):
coef3.append(-C[k,i,j])
vars3=[]
vars3.append(s[k])
for i in range(I):
vars3.append(x[i])
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=vars2,val=coef2)],senses=["L"], rhs=[M2], names=['bound_s1[%s,%s]' %(k,j)])[0]
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=vars3, val=coef3)], senses=["G"], rhs=[0], names=['bound_s2[%s,%s]' %(k,j)])[0]
model._data = x, q, s, theta
# model.write("D2HAHA.lp")
return model
G_Separation问题
def SeparationProblem_q_C_norm(p,q,theta,I,J,K,R,C):
model=cplex.Cplex()
alpha,beta,gamma,lamb={},{},{},{}
for k in range(K):
for i in range(I):
alpha[k,i]=model.variables.add(obj=[0], lb=[-cplex.infinity], ub=[cplex.infinity], types=["C"], names=["alpha[%s,%s]"%(k,i)])[0]
for k in range(K):
for j in range(J):
beta[k,j]=model.variables.add(obj=[q[k,j]], lb=[-cplex.infinity], ub=[cplex.infinity], types=["C"], names=["beta[%s,%s]"%(k,j)])[0]
for k in range(K):
for l in range(J):
for j in range(J):
gamma[k,l,j]=model.variables.add(obj=[0], lb=[0], ub=[cplex.infinity], types=["C"], names=["gamma[%s,%s,%s]"%(k,l,j)])[0]
for k in range(K):
lamb[k]=model.variables.add(obj=[p[k]*theta[k]], lb=[0], ub=[cplex.infinity], types=["C"], names=["lambda[%s]"%(k)])[0]
model.objective.set_sense(model.objective.sense.minimize)
#Restricciones
for k in range(K):
for i in range(I):
for j in range(J):
var=[]
var.append(alpha[k,i])
var.append(beta[k,j])
var.apend(lamb[k])
for l in range(J):
if l!=j:
var.append(gamma[k,l,j])
coef=[]
coef.append(1)
coef.append(1)
coef.append(p[k]*R[k,i,j])
for l in range(J):
if l!=j:
coef.append(C[k,i,l]-C[k,i,j])
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=var,val=coef)], senses=["G"], rhs=[0], names=['1era[%s,%s,%s]'%(k,i,j)])[0]
for i in range(I):
var=[]
for k in range(K):
var.append(alpha[k,i])
coef=[]
for k in range(K):
coef.append(-1)
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=var,val=coef)], senses=["G"], rhs=[0], names=['2a[%s]'%(i)])[0]
var=[]
for k in range(K):
var.append(lamb[k])
for j in range(J):
for k in range(K):
var.append(beta[k,j])
coef=[]
for k in range(K):
coef.append(-1)
for j in range(J):
for k in range(K):
coef.append(1)
model.linear_constraints.add(lin_expr=[cplex.SparsePair(ind=var,val=coef)], senses=["E"], rhs=[-1], names=['norm'])[0]
model._data = alpha,beta,gamma,lamb
return model
我的实例还可以，并且可以与其他算法一起运行而没有问题。
答案 0 :(得分：1)
调试回调错误的最好方法是将回调代码包装到try/except块中，从中打印出实际的异常，然后重新引发它：
try:
# callback code here
...
except:
print(sys.exc_info()[0])
raise
这将告诉您到底出了什么问题。