Originally posted by: LJXue

I am using cplex 12.3 with python 2.7 under Windows 7. Everything is in 64 bit.

I have the following issue which is suspicious to be a bug. In the example, I create a problem with 3 variables, and add another variable later on. Although I specify the variable name, the variable type does not match the sequence I specify.

1. create the model

A = [http://-1.0, 1.0, 1.0, http://1.0, -3.0, 1.0, http://0.0, 1.0, 0.0]
rhs = http://20.0, 30.0, 0.0
cSense =
nCon = 3
var_names =

obj = http://1.0, 2.0, 3.0
ub = http://40.0, cplex.infinity, cplex.infinity
prob.variables.add(obj = obj, ub = ub, types = 'CCC', names = var_names)

const_expr = []
for i in range(nCon):
temp = [var_names, Ai]
const_expr.append(temp)

prob.linear_constraints.add(lin_expr = const_expr, senses = cSense, rhs = rhs)

qmat = [[0, 1, http://-33.0, 6.0],
[0, 1, 2, http:// 6.0, -22.0, 11.5],
[1, 2, http:// 11.5, -11.0]]

prob.objective.set_quadratic(qmat)
print(prob.variables.get_names())
print(prob.variables.get_types())

1. output:

#
#

1. add an additional variable

prob.variables.add(obj = http://1.0, lb = http://0.0, ub = http://3.0, types = 'I',
names = ,
columns = [[0, 2, http://10.0, -3.5]])
print(prob.variables.get_names())
print(prob.variables.get_types())
#output
#

1. # this is not right, x4 should be integer variable

Currently, a workaround I use is not to specify the variable type in add variable statement
and change it via

prob.variables.set_types()

I wonder if there is any mistake in my statements, or it is a bug. Thank you very much.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: LJXue

sorry, seems there are some formatting issues.

#create the model
 
import cplex
 
prob = cplex.Cplex()
prob.set_problem_name('Example')
prob.objective.set_sense(prob.objective.sense.maximize)
 
 
A = [[-1.0, 1.0, 1.0], [1.0, -3.0, 1.0], [0.0, 1.0, 0.0]]
rhs = [20.0, 30.0, 0.0]
cSense = ['LLE']
nCon = 3
var_names = ['x' + str(i + 1) for i in range(nCon)]
 
obj = [1.0, 2.0, 3.0]
ub = [40.0, cplex.infinity, cplex.infinity]
prob.variables.add(obj = obj, ub = ub, types = 'CCC', names = var_names)
 
const_expr = []
for i in range(nCon):
    temp = [var_names, A[i]]
    const_expr.append(temp)
 
prob.linear_constraints.add(lin_expr = const_expr, senses = cSense, rhs = rhs)
 
qmat = [[[0, 1], [-33.0, 6.0]],
        [[0, 1, 2], [ 6.0, -22.0, 11.5]],
        [[1, 2], [ 11.5, -11.0]]]
 
prob.objective.set_quadratic(qmat)
 
print(prob.variables.get_names())
print(prob.variables.get_types())
#output
#['x1', 'x2', 'x3']
#['C', 'C', 'C']
 
# add another variable
prob.variables.add(obj = [1.0], lb = [0.0], ub = [3.0], types = 'I',
                   names = ['x4'],
                   columns = [[[0, 2], [10.0, -3.5]]])
print(prob.variables.get_names())
print(prob.variables.get_types())
#output
#['x1', 'x2', 'x3', 'x4']
#['I', 'C', 'C', 'C'] # this is not right, x4 should be integer variable

Currently, a workaround I use is not to specify the variable type in add variable statement
and change it via

prob.variables.set_types()

I wonder if there is any mistake in my statements, or it is a bug. Thank you very much.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Hello,

Thank you for your report, this is indeed a bug in the Python interface code. When variables are added to the problem without new columns in the constraint matrix, this bug is not triggered.

If the workaround that you have identified is not adequate, please contact me at pstarhill (at) us (dot) ibm (dot) com.

Philip Starhill
CPLEX Research Engineer
#CPLEXOptimizers
#DecisionOptimization