Originally posted by: SystemAdmin

Hey,

I am using Cplex (Version 12,3) together with Matlab. I wanted to try the new functionality to solve nonconvex quadratic problems. Although I set the solutiontarget to 2 I still receive the CPLEX Error " 5002: Q in objective is not positive semi-definite." Do I need to set any other properties, or am I missing something else? Here is a small testconfiguration I tried in Matlab with the target function X2*X1 + 10*X3*X1 (the problem I want to solve in the long run is much larger)

cplex = Cplex(
'mipex1'); cplex.Model.sense = 
'minimize';   cplex.Model.obj   =  [0,0,0]; cplex.Model.lb    =  [0,0,0]; cplex.Model.ub    =  [inf,inf,inf]; cplex.Model.ctype =  
'CC'; cplex.Model.A     =  [ 1, 1 ,0 ]; cplex.Model.lhs   =  [10]; cplex.Model.rhs   =  [inf]; cplex.Model.Q     =  [0   , 0.5 , 5 0.5 , 0   , 0 5   , 0   , 0];   cplex.Param.solutiontarget.Cur = 2;   cplex.solve;

With the cplex result:

Tried aggregator 1 time. MIQP Presolve eliminated 0 rows and 1 columns. Number of nonzeros in lower triangle of Q = 1 Using Approximate Minimum Degree ordering Total time 

for automatic ordering = 0.00 sec. Summary statistics 

for factor of Q: Rows in Factor            = 2 Integer space required    = 2 Total non-zeros in factor = 3 Total FP ops to factor    = 5 Presolve time =    0.01 sec. Error using cplexlink123 *CPLEX Error  5002: Q in objective is not positive semi-definite.*

It would be awsome if anyone had an idea how to solve my problem.
Thanks in advance,

Felix
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: John Cui

Your problem is a QP, but a MIQP, right? then please don't set cplex.Model.ctype.

And also, CPLEX only can solve non-convex QP with solutiontarget to 2 currently in CPLEX 12.3, CPLEX12.4.

John Cui
#CPLEXOptimizers
#DecisionOptimization